All Texts
Blog Post
Reflective

The First Time I Used AI

📂 Technology, Work, And Everyday Life 🎭 What An Ordinary First Encounter With AI Reveals About Writing, Thinking, And What We Want From A Tool ⏱ 20–55 min
About this text
🎯 Learning objectives
  • Students can read and understand a first-person reflective text about a contemporary topic.
  • Students can use past tenses to describe a first experience — past simple, past continuous, and at higher levels past perfect.
  • Students can describe how a tool, app, or website is used in basic terms.
  • Students can talk about the difference between writing for yourself and writing for someone else.
  • Students can discuss how AI tools are being used in their own country, school, or workplace.
  • Students can write a short reflective piece about a tool or technology they have started using.
  • Students can express measured, mature opinions about technology, authorship, and what they want from new tools.
💡 Ideas for using this in a lesson
  • Pair work: students interview each other about the first time they used a particular tool — a phone, a translation app, AI, a particular website.
  • Vocabulary mapping: students collect every word in the text related to writing and editing — draft, send, prompt, sentence — and group them.
  • Past-tense practice (B1+): students write five sentences about a first experience with a new tool, using past simple and past perfect.
  • Cultural sharing: 'How are AI tools being used in your country, school, or workplace? Are they welcomed, debated, or restricted?' Students share in small groups.
  • Sequencing activity (A1/A2): cut the steps of the writer's first AI use into strips. Students put them in order.
  • Writing task: students write a short reflective piece at their level about the first time they used a particular technology and what they learned.
  • Discussion (B1+): 'Is using AI to help write a personal message a kind of cheating, or just a tool — like a dictionary or a spell-checker?'
  • Role-play (B2+): a student plays someone defending their use of AI for a personal task; another student plays a sceptical relative or friend. Practise polite disagreement.
  • Comparison task (B2+): students compare the A2 and C1 versions and identify what is added at the higher level — particularly the awareness of authenticity, voice, and what it means to outsource thought.
  • Reflective writing (C1+): 'A tool I have started using that has changed how I think, not just what I do.' Personal essay.
  • Critical reading (C2): students identify the moments when the writer's argument turns on itself — where the same observation could be read as praise or critique. Why does the writer keep both readings on the page?
🏷️ Context
PairworkGroupworkPersonal TopicReflectiveNarrativeEveryday VocabularyTechnologyModern LifeDiscussionRelevant 2026Contemporary
📦 Materials needed
Paper And Pen
⚠️ The topic is current and highly relevant for many students, but it carries some real complications. AI tools are now used widely in education, and many students will have used them — sometimes in ways their schools or universities have rules about. The text deliberately does not take a strong position on whether AI use is good or bad, and teachers should follow this lead. Students may have strong feelings — excitement, discomfort, anxiety about jobs, scepticism — and all of these should be welcomed. Some students may come from countries or workplaces where AI access is restricted; others may use these tools daily. The lesson should not assume access. There is also class to think about: in some contexts, AI tools are seen as tools of efficiency for professionals; in others, as a workaround for people learning the language; in others, as a threat to writing as a discipline. The text at higher levels addresses these tensions. Students who are themselves writers, journalists, translators, or teachers may have particular views, and these are valuable. Teachers should make space for ambivalence rather than push students toward a settled position.
⏱ Duration by level
A1
20 min
A2
25 min
B1
35 min
B2
45 min
C1
50 min
C2
55 min
🎚️ Differentiation tip
For A1 and A2, focus on the simple narrative — opened the page, wrote a question, read the answer — and on basic vocabulary of writing and asking. Past simple is the main grammar focus. For B1, introduce the small reflective layer the writer adds: what surprised them, what was harder than expected, what was not what they had pictured. For B2, the focus shifts to voice — slightly self-aware, careful not to overclaim, willing to register both relief and unease. For C1 and C2, the post becomes a meditation on what writing is for, what gets lost when a difficult sentence is outsourced, and how the easy availability of a tool can change the shape of the task. The C2 in particular is openly aware of itself as an essay written, in part, in the same era as the tool it describes — and asks the harder question of what kind of authorship is even available now.
🌍 Cultural note
AI tools — chatbots, writing assistants, translation systems — have, in the past few years, moved from specialist novelty to everyday infrastructure in many countries, but their reach is uneven. Some students will be using them daily for school or work; others will have only heard about them; others may be in contexts where access is restricted by language, region, government policy, or institutional rules. Attitudes also vary widely. In some workplaces, AI use is mandatory; in others, it is forbidden in certain tasks (for example, in some legal, medical, or academic settings). For language learners specifically, AI tools can be both a tremendous help (instant feedback, conversation practice, translation) and a complicated one (over-reliance, getting answers without learning). Teachers should welcome students' actual experiences — including students who have not used these tools and have no particular wish to — without ranking them. The writer's particular use case in this text (writing a difficult personal email) is one possibility among many; readers from translation, education, healthcare, and the arts will all have different stories.
Beginner
Intermediate
Advanced
Duration: 20 min 🎯 Focus: Past simple, time markers (last month, the next day), basic possessives (my, his), simple coordinators (and, but), common adjectives (difficult, sorry, okay, good).
Before You Read / Listen
  • Q1Do you use a computer or a phone every day?
  • Q2Have you used AI? What do you know about it?
  • Q3Is it easy or difficult for you to write an email?
  • Q4When was the last time you said sorry to someone?
  • Q5Do you like new technology? Yes, no, or sometimes?
The Text
Last month, I used AI for the first time. I am forty-two years old. I do not love new technology.
I had to write a difficult email. The email was for my brother. We had a small argument. I was sorry.
I sat at my desk for an hour. I could not start. The first sentence was too hard.
My friend said, 'Try AI. It can help you.' I opened a website on my computer.
I wrote one sentence 'Help me write an email to my brother. I want to say sorry.' I clicked a button.
The AI wrote an email for me. The email was good. But it was not me.
I read it three times. Then I wrote a new email, in my own words. The AI helped me start.
I sent the email to my brother. He called me the next day. We are okay now.
AI is a tool. It is not magic. But it helped me at a difficult moment. I am not sure how I feel about it.
Key Vocabulary
AI noun
a computer program that can write, answer questions, and help with tasks
"I used AI for the first time."
email noun
a message you send on the internet
"I had to write a difficult email."
argument noun
when two people disagree and are angry
"We had a small argument."
sorry adjective
feeling bad about something you did
"I was sorry."
click verb
to press a button on a computer or phone
"I clicked a button."
tool noun
a thing that helps you do work
"AI is a tool."
magic noun
something not real, like in a story
"AI is not magic."
send verb
to give a message to someone, by post or computer
"I sent the email."
moment noun
a very short time
"It helped me at a difficult moment."
Questions
Comprehension
  • How old is the writer?
    Answer
    Forty-two.
  • Why did the writer need to write an email?
    Answer
    Because the writer had a small argument with their brother. The writer was sorry.
  • Who told the writer to try AI?
    Answer
    The writer's friend.
  • What did the writer ask the AI?
    Answer
    To help write an email to the brother to say sorry.
  • Did the writer send the AI's email?
    Answer
    No. The writer wrote a new email in their own words. The AI helped the writer start.
  • What happened the next day?
    Answer
    The brother called the writer. They are okay now.
Vocabulary
  • What is 'AI'?
    Answer
    A computer program that can write, answer questions, and help with tasks.
  • What is an 'argument'?
    Answer
    When two people disagree and are angry.
  • What is a 'tool'?
    Answer
    A thing that helps you do work. AI is a tool.
Personal
  • Have you used AI? Or do you know someone who has?
    Teacher guidance
    Students' own answers. Welcome simple yes/no answers and short stories. Some students will have used it a lot; others not at all. Both are fine. Help with vocabulary they don't yet know in English.
Discussion
  • Is it good or bad to use a computer to help you write?
    Discussion prompts
    Prompts for discussion: good — saves time, helps when you are stuck, helps with new languages, useful for work. Bad — you don't learn, the computer makes mistakes, the writing is not really yours, you can become lazy. Both are honest answers. Many students will have done both.
Writing Task
Prompt
Write 5–7 sentences about the first time you used a new technology — a phone, a computer, a website, or AI. Use the past simple. Say what it was, when it was, and how you felt.
Model Answer

The first time I used a smartphone, I was thirty years old. My son gave me his old phone. I was scared. The screen was small. I did not know how to send a message. My son helped me. After a week, it was easier. Now I use it every day.

Activities
  • Listen and repeat: the teacher reads the story slowly. Students repeat each sentence.
  • Vocabulary picture: draw a person at a computer. Label five things: computer, screen, keyboard, button, person.
  • Pair work: in pairs, ask and answer five questions about the story.
  • Past tense practice: change five sentences from past to present (e.g. 'I used AI' → 'I use AI').
  • Sequence: the teacher gives nine sentences from the story in the wrong order. Students put them in the right order.
  • Match: match nine words to nine pictures (computer, email, button, brother, friend, screen, message, tool, AI).
  • Write: write 5–7 sentences about the first time you used a new technology.
  • Speak: tell your partner about one technology you use every day.
Duration: 25 min 🎯 Focus: Past simple, past continuous, present perfect ('I have been thinking'), connectors (but, and, because, when), modals (can, would), 'I am not sure' for hedging.
Before You Read / Listen
  • Q1Have you ever had a difficult email or message to write? What was it about?
  • Q2When you do not know how to start something, what do you do — wait, ask a friend, use the computer?
  • Q3Have you used AI? If yes, what for? If no, would you try it?
  • Q4Is there a person in your family that you have not spoken to for a while? You don't have to share who.
  • Q5Do you think a computer can really help with feelings, or only with information?
The Text
Last month, I used AI for the first time. I am forty-two years old, and I am not really a 'technology person'. I have a phone and a computer for work, but that is all.
I needed to write a difficult email. It was to my younger brother. We had a small argument in the summer, and we had not really spoken since. I was sorry, but I did not know how to start. I sat at my desk for almost an hour, looking at an empty page.
A friend at work, who uses these things every day, said, 'Just try the AI. It will not write the email for you, but it can help you start.' I was tired, and the empty page was not getting smaller.
I opened a website. I typed 'Help me write an email to my brother. We had an argument. I want to say sorry, but I do not want to sound cold.' I clicked the button. The reply came back in two seconds.
The AI wrote a full email. It was polite. It said the right kinds of things. I read it three times. And then I knew I could not send it. The email was good — but it was not me. My brother would know.
What happened next was a surprise. I started writing my own email. I just used the start. The AI had given me one sentence — 'There is something I have been wanting to say' — and I built my email from there. The rest was mine.
I sent the email at midnight. My brother called me the next morning. We talked for an hour. We are okay now.
AI is, as my friend said, a tool. It did not write the email for me. But it broke the silence in my head. I had been afraid of the empty page. Once I had a starting point, I knew what I really wanted to say.
I am not sure how I feel about this. I am glad I sent the email. But the email had a small piece of someone else's writing inside it, and even though my brother does not know, I know. That is something I am still thinking about.
Key Vocabulary
technology person noun phrase
someone who likes and uses new technology a lot
"I am not really a 'technology person'."
younger brother noun phrase
a brother who is born after you
"I have a younger brother."
empty page noun phrase
a page with nothing on it
"I sat looking at an empty page."
type verb
to write using the keys of a computer or phone
"I typed a question."
polite adjective
with good manners; respectful
"The email was polite."
cold (in writing) adjective
without warmth; not friendly
"I did not want to sound cold."
starting point noun phrase
the place where you begin
"Once I had a starting point, I knew what to say."
break (the silence) verb
to end a quiet, difficult time
"The AI broke the silence in my head."
be afraid of verb phrase
to feel fear
"I had been afraid of the empty page."
even though conjunction
(connector) although; even when
"Even though my brother does not know, I know."
Questions
Comprehension
  • Why did the writer need to write the email?
    Answer
    Because they had a small argument with their younger brother in the summer at their parents' house, and they had not really spoken since. The writer was sorry.
  • How long did the writer sit looking at the empty page?
    Answer
    Almost an hour.
  • What did the writer's friend say about AI?
    Answer
    That it is just a tool, that it will not write the email for the writer, but it can help them start.
  • Why did the writer not send the AI's email?
    Answer
    Because the email was good but it was not the writer. The writer thought their brother would know it was not in their voice.
  • What did the writer use from the AI's email?
    Answer
    Just the first sentence: 'There is something I have been wanting to say for a few months now.' The rest of the email was the writer's own.
  • What happened the morning after the writer sent the email?
    Answer
    The brother called the writer. They talked for an hour. They are okay now.
  • What is the writer's view of AI at the end of the post?
    Answer
    That AI is a tool. It did not write the email, but it broke the silence in the writer's head. Once the writer had a starting point, they knew what they really wanted to say.
Vocabulary
  • What does 'a technology person' mean?
    Answer
    Someone who likes and uses new technology a lot. The writer says they are not one.
  • What does 'cold' mean in 'I did not want to sound cold'?
    Answer
    Without warmth; not friendly. The writer was worried that their email might sound distant or unfriendly to their brother.
  • What does 'break the silence' mean here?
    Answer
    To end a quiet, difficult time. The 'silence' was in the writer's head — they had been unable to start writing. The AI broke it.
Inference
  • Why did the writer say 'the empty page was not getting smaller'?
    Suggested interpretation
    Because no matter how long the writer waited, the page stayed the same size — empty. It was not going to fix itself. The writer is suggesting that waiting was not working, so they had to try something else.
  • Why is the writer 'still thinking about' the email?
    Suggested interpretation
    Because the email had a small piece of someone else's writing inside it. The brother does not know this, but the writer does. The writer is trying to work out whether this matters and what it means about who actually wrote the email.
Personal
  • Have you ever waited a long time before writing or saying something difficult? What helped you start?
    Teacher guidance
    Students' own answers. Listen for honest answers — a friend pushed them, time helped, they wrote it down first, they did it at night when no one was watching. Common patterns: 'I waited too long', 'A friend told me to just do it'. Validate without prying.
  • Have you used a tool to help you write or speak in English — a translator, a spell-checker, a dictionary, AI? Did it help, or did you become too dependent on it?
    Teacher guidance
    Students' own answers. Many students will have stories. Listen for both sides — tools are helpful, but over-use can stop learning. Common patterns: 'I use it sometimes but try not to depend on it', 'It helped me but my friend says I should write more without it'. Validate the complexity.
Discussion
  • Is it 'cheating' to use AI to help write a personal message? Or is it just a tool, like a dictionary?
    Discussion prompts
    Prompts for discussion: cheating — the words are not really yours; the person reading the message thinks they came from you, so it is dishonest. Just a tool — we all use spell-checkers, dictionaries, and friends to help us write; AI is the same. Real answer: depends on how it is used. Using it to start, like the writer in the story, is different from using it to write the whole message. Most students will recognise both views.
Writing Task
Prompt
Write 8–10 sentences about a time you had to do something difficult and got help (from a person, a book, a website, or a tool). Say what was difficult, who or what helped you, and how you felt afterwards. Use past simple and past continuous.
Model Answer

Two years ago, I had to give a speech at my sister's wedding. I had never spoken to a big group of people before. I sat at home for many evenings and could not start. My older sister called me one night. She said, 'Just write the first sentence. The rest will come.' I wrote one sentence — 'I have known my sister for thirty years' — and after that, the speech started to come. On the wedding day, I read my speech and people laughed in the right places. I was tired but happy. My sister said it was a good speech.

Activities
  • Read and discuss: in pairs, students read the post and identify the moment the writer felt better.
  • Vocabulary practice: students put ten new words into sentences about their own experience.
  • Past tense practice: students write five sentences about a difficult task they did.
  • Pair interview: student A asks student B about a tool they have used recently. Then swap.
  • Sequence: cut the steps of the writer's evening into strips (sat at desk, opened website, typed question, read AI's email, wrote own email, sent at midnight, brother called). Put them in order.
  • Discussion in groups of three: 'Is using AI to help with writing okay, or not? When?' Students share their honest opinions.
  • Write: 8–10 sentences about a time you got help with something difficult.
  • Speak: in pairs, share one thing you would like help with — and what kind of help you would want.
  • Reflect: 'Something I would like to say to someone but have not yet' — students write three sentences. They do not have to share.
Duration: 35 min 🎯 Focus: Past simple, past continuous, past perfect ('I had been sitting'), present perfect ('I have been thinking'), reflective discourse, hedging (it does not feel like nothing, either), connectors of contrast (but, however), reported speech, sensory description.
Before You Read / Listen
  • Q1Have you ever waited too long to send a message and watched it become harder, not easier?
  • Q2Have you ever used AI? If yes, what for? If no, would you?
  • Q3When you are stuck on something — a difficult email, a difficult choice — what helps you start?
  • Q4Is there a difference, for you, between getting help from a friend and getting help from a tool?
  • Q5Have you ever had a 'first experience with a technology' that you remember clearly?
The Text
I want to tell you about the first time I used AI. It happened last month. I am forty-two years old, and I would not normally write about something like this — but the experience surprised me, and I am still thinking about it.
I should explain the situation. I needed to write a difficult email to my younger brother. We had argued at our parents' house in the summer, about something that turned out to be smaller than it had felt at the time, and we had not really spoken since. Three months had passed. I was sorry. But I did not know how to start the email, and the longer I waited, the harder starting became.
I had been sitting at my desk on a Sunday evening, looking at an empty document, for the better part of an hour. A friend who had come round for tea had said, slightly impatiently, 'Have you tried just opening one of those AI things? It is not going to write the email for you. It might help you stop staring at the wall.' I had said something dismissive at the time. I am, by temperament, suspicious of new technology. But I was now an hour into a Sunday evening with no email written, and the page was not going to write itself.
I opened a website I had heard about but never used. There was a small white box at the top of the screen. I typed, with more honesty than I would have used with most humans, that I had had an argument with my brother and not spoken to him for three months and that I needed to write an email saying sorry but did not want to sound cold. I clicked the button. Two seconds later, an email appeared on my screen.
I want to be careful about what I say next, because I know this is the part where the writer either declares the technology a miracle or dismisses it as nothing. The truth was somewhere stranger. The email was, in a technical sense, a perfectly serviceable email. It used my brother's name. It said the right kinds of things. It was the kind of email I might have written if I had been writing it for somebody else, on a project, with no real feelings involved. I read it three times. And I knew I could not send it. The email was good and it was not mine.
What happened next is the part I have been trying to describe to people, and have not yet got right. I started, almost without thinking about it, to write my own email. The AI's words were in the back of my head, but I did not use them, except for the opening — 'There is something I have been wanting to say for a few months now' — which I kept because it sounded like the way I might actually start. The rest of the email was, sentence by sentence, mine.
I sent the email just before midnight. My brother called me the next morning. We talked for an hour. We are, as people sometimes say, okay now — which is a small word for a complicated thing.
I have been trying to work out what the AI actually did for me. The temptation is to say it did either everything or nothing. Neither is right. What it did was simpler. It put a sentence on the page where there had been no sentence. It broke the empty-page silence I had been losing to for an hour. Once there was a sentence on the page — even a sentence I did not want to keep — I could see what I disagreed with, what I would have said differently. I had not known, before that evening, how much of the difficulty of writing comes from the empty page itself.
I should also say, because I want this to be honest, that the email had a small piece of someone else's writing inside it, and that my brother does not know this. It does not feel like cheating, exactly. But it does not feel like nothing, either. I sometimes wonder whether the email I would eventually have written, on my own, would have been different — and whether the difference would have been important.
I have, since that Sunday, used the same tool twice more — once to start a work email I had been avoiding, and once to draft a message I did not, in the end, send. The pattern is becoming familiar. It does not write what I want to say. It produces something close enough to what I might say that I can tell, by reaction, what I actually mean. That is, in some unfamiliar way, useful. It is also, in some unfamiliar way, slightly strange. I am keeping both impressions in mind.
Key Vocabulary
earn (a piece of writing) verb
(figurative) to deserve being written about
"It has earned a piece of writing."
by temperament phrase
by natural character or disposition
"I am, by temperament, suspicious of new technology."
dismissive adjective
treating something as not important
"I had said something dismissive at the time."
prompt (in technology) noun
a question or request asking for input
"A friendly little prompt asking what I needed."
serviceable adjective
good enough for the purpose; functional
"A perfectly serviceable email."
acknowledgement noun
the act of recognising or accepting something
"An acknowledgement that some time had passed."
stare (at) verb
to look at something for a long time without moving
"Stop staring at the wall."
strange (in a feeling) adjective
different from what is expected, slightly uncomfortable
"I did not want to sound cold or strange."
miracle noun
a wonderful, almost impossible thing
"Either declares the technology a miracle or dismisses it as nothing."
draft (a message) verb
to write a first version of something
"Once to draft a message I did not, in the end, send."
by reaction phrase
by responding to something already in front of you
"I can tell, by reaction, what I actually mean."
Questions
Comprehension
  • Why did the writer need to write the email, and how long had it been since the argument?
    Answer
    Because they had argued with their younger brother at their parents' house in the summer, three months earlier, and had not really spoken since. The argument had become harder to talk about, not easier.
  • What did the writer's friend say about AI, and what was the writer's first reaction?
    Answer
    The friend said, slightly impatiently, that the AI is not going to write the email but might help the writer stop staring at the wall. The writer said something dismissive at the time, being suspicious of new technology by temperament and not wanting to admit they needed help with a sentence to their own brother.
  • What did the writer think of the email the AI produced?
    Answer
    It was a perfectly serviceable email — it used the brother's name, opened with an acknowledgement that some time had passed, and said the right kinds of things. But it was the kind of email the writer might have written for somebody else, on a project, in an office, with no real feelings involved. The writer could not send it.
  • What part of the AI's email did the writer keep, and why?
    Answer
    The opening line, 'There is something I have been wanting to say for a few months now', because it sounded like the way the writer might actually start.
  • What does the writer say the AI actually did for them?
    Answer
    It put a sentence on the page where there had been no sentence. It broke the empty-page silence the writer had been losing to for an hour. Once there was a sentence on the page — even one the writer didn't want to keep — they could see what they disagreed with, what they would have said differently, and what was missing.
  • What pattern has the writer noticed in their later uses of the tool?
    Answer
    It does not write what the writer wants to say. It produces something close enough to what the writer might say that the writer can tell, by reaction, what they actually mean.
Vocabulary
  • What does 'serviceable' mean, and why does the writer use it for the AI's email?
    Answer
    Good enough for the purpose; functional. The writer uses it to mean the email worked technically — it covered the right ground — without being exactly right. It is a slightly cool word that praises the email's competence while withholding warmer praise.
  • What does 'dismissive' mean, and what does it tell us about the writer's first reaction to the friend's suggestion?
    Answer
    Treating something as not important. The writer brushed off the friend's suggestion at first — they didn't take it seriously. The word admits the writer's initial resistance honestly.
  • What does the writer mean by 'I can tell, by reaction, what I actually mean'?
    Answer
    That seeing something in front of them — even something they don't fully agree with — helps them notice what they would say differently. They learn what they think by reacting to a draft, not by trying to produce one from scratch. It is a particular kind of thinking-through-response.
Inference
  • Why does the writer say 'I had not known, before that evening, how much of the difficulty of writing comes from the empty page itself'?
    Suggested interpretation
    Because the writer had assumed the difficulty was about not knowing what to say. The evening showed them that some of the difficulty was a separate thing — the resistance produced by an empty document. Once there was anything on the page, the resistance dropped, even though the writer ended up writing their own words. The reasoning: the writer is naming a discovery they didn't know they were making.
  • Why does the writer say the email 'does not feel like cheating, exactly' — but also 'does not feel like nothing, either'?
    Suggested interpretation
    Because cheating implies passing off someone else's work as your own with intent to deceive, and the writer's email is mostly their own work. But the small piece they kept (the opening line) is not theirs, and the brother does not know — so it is not entirely innocent either. The writer is honestly sitting between two judgements rather than choosing one.
  • What is the writer's view of how AI writing about AI tends to go, and how do they try to avoid it?
    Suggested interpretation
    The writer says AI essays usually either declare the technology a miracle or dismiss it as nothing. They try to avoid both by saying 'the truth was somewhere stranger' and describing the experience precisely — what the AI did, what it didn't, how the writer felt at each step — without reaching for either large category.
Discussion
  • Is the writer right that part of the difficulty of writing comes from the empty page itself, separate from not knowing what to say? Where else in life does this kind of starting-difficulty appear?
    Discussion prompts
    Prompts for discussion: candidates beyond writing — the first phone call after a long silence, the first time at the gym, opening a difficult conversation, starting a new project, sitting down to study. The starting friction is often larger than the actual task. Real answer: most students will recognise the experience. The interesting follow-up is what helps each person break it.
  • Is using AI to help write a personal message different from using a friend's advice, a translation app, or a phrase book? Where is the line?
    Discussion prompts
    Prompts for discussion: same — they are all forms of help; people have always asked for advice on difficult letters and emails. Different — friends know you, AI does not; friends share responsibility for the message, AI does not; AI patterns may shape your writing in ways you don't notice. Real answer: depends on how the help is used. The writer's use (a starting line, then their own work) sits closer to a friend's advice than to writing a whole message.
  • The writer says they used AI for two more things — one work email, and one message they did not send. Why might it be useful to draft a message you do not send?
    Discussion prompts
    Prompts for discussion: it lets you see what you would have said, without committing to saying it; it gives the feelings a place to go; it can change what you do say later, even if you don't send the original. Some people find writing-without-sending genuinely valuable. Others find it leaves them more entangled with a situation they might have left alone. Real answer: it depends on whether the writing is a step toward action or a substitute for it.
Personal
  • Have you ever had a difficult message you needed to send, and waited too long? What helped you finally write it?
    Teacher guidance
    Students' own answers. Be careful — these questions can touch on family disagreements, lost friendships, difficult work situations. Listen first. Common patterns: 'A friend pushed me', 'I wrote it down at night', 'I just had to do it'. Don't pry. Welcome students who say they are still waiting.
  • Have you used AI for something? If yes, what was useful, and what felt strange? If no, what is stopping you?
    Teacher guidance
    Students' own answers. Many will have used it; some won't. Welcome both. Common patterns: 'I use it for work but I don't trust it for personal things', 'I tried it once and it made me feel weird', 'I haven't tried it because...'. Validate the variety.
Writing Task
Prompt
Write a 250–400-word reflective post about the first time you used a particular tool, technology, or service that you were unsure about. Be honest about your initial reaction, what surprised you, and what you are still thinking about. Use past tenses confidently.
Model Answer

Two years ago, I started using a translation app on my phone for the first time. I had moved to a new country for work, and I knew about thirty words of the local language. The first weeks were exhausting. I had been carrying a small phrase book that I had bought at the airport, and I had been very proud of myself for using it, even when the result was usually that the other person looked at me kindly and switched to English. Then a colleague at work showed me the translation app. I was, at first, annoyed by it. It felt like a kind of giving up. The phrase book had been mine — I had circled the pages I used most, written notes in the margins, learned to say 'Excuse me, where is the post office?' with what I had imagined was a careful accent. The app, by comparison, was anonymous. It just spoke for me. The first time I used it was at a small bakery near my flat. I needed to ask whether the bread had nuts in it. I typed the question, the phone said the question out loud, and the woman behind the counter answered, naturally, like I was a normal customer. I bought the bread and walked home and sat on a bench for a few minutes feeling quite strange. The transaction had been so normal. I had been treated as someone who wanted a piece of bread, not as a foreigner performing a small effortful play of belonging. I have used the app most days since. It has, in fairness, helped me learn the language faster — because I now have actual conversations rather than careful broken ones. But I still miss the phrase book a little. The app is more useful and more anonymous, and I am not sure that those two things are entirely separable.

Activities
  • Pre-reading: in pairs, students answer the intro questions and choose one to share with the class.
  • Vocabulary in context: students underline ten new words and explain what each adds to the meaning of the sentence.
  • Past tense practice: students write five sentences about a first experience with a tool or technology, using past simple and past perfect.
  • Discussion in groups: 'When does help with writing become a problem? Where is the line?' Surface arguments on both sides.
  • Voice analysis: students find three sentences where the writer is being honest about being unsure, and discuss why this works.
  • Pair interview: students interview each other about a tool they have started using. Practise active listening — what makes the story interesting?
  • Writing: students draft the 250–400-word reflective post.
  • Peer feedback: in pairs, students read each other's posts and write one warm comment and one practical question.
Duration: 45 min 🎯 Focus: Cohesion across long paragraphs, hedging and qualification, mature reflective register, nominalisation (resistance, friction, mechanical), wry voice and self-awareness about the genre, complex sentence structures with embedded subclauses, present perfect alongside past simple.
Before You Read / Listen
  • Q1What is a piece of writing about technology you have read recently that you found thoughtful, and what made it that way?
  • Q2Have you noticed parts of your work or thinking that turn out to be 'mechanical' — separable from the parts that feel uniquely yours?
  • Q3Is there a difference, in your view, between 'difficulty of starting' and 'difficulty of saying'? Where do you most experience each?
  • Q4Have you ever taken a position on a new technology that, looking back, was more about temperament than about the technology itself?
  • Q5If a tool can do part of what you used to do, are you free to do something more interesting, or are you slowly being made smaller? Both? When?
The Text
I want to tell you about the first time I used AI, with the small caveat that the phrase 'the first time I used AI' has, in 2026, the same slightly worn quality as 'the first time I went on the internet' had in 2002 — universal enough that the writer of the essay should be aware of how universal it is, and rare enough as a deliberately chosen subject that some honest reflection on it might still be worth producing. I am forty-two years old. I am, by temperament, suspicious of new technology, and I have spent the past year resisting both the people who are excited about these tools and the people who are sure they will end civilisation. The truth, as I have come to it, is somewhere stranger and smaller than either, and I would like to try to put it down.
The situation, briefly. I needed to write a difficult email to my younger brother. We had argued at our parents' house in the summer about something that, in retrospect, was smaller than it had felt, and we had not really spoken in three months. The argument had become, in the way these things do, harder to address rather than easier. I was sorry. But I had not been able to start the email, and the not-starting had become a small ongoing fact of my life — a thing I would notice on the walk to work, and at the kitchen sink, and in the soft moment between waking up and remembering what was happening that day.
On a Sunday evening, for reasons I cannot now reconstruct except as the gradual erosion of resistance, I opened a website I had heard about but never used. I typed, with more honesty than I would have used with most humans, the gist of the situation: I had had an argument with my brother in the summer; we had not spoken since; I wanted to write an email saying sorry; I did not want it to sound cold or strange. I clicked the button. The reply came in two seconds.
I want to be careful about how I describe what came back, because the genre of writing about this experience tends to do one of two things — either to declare the technology a miracle and write breathlessly about its capabilities, or to dismiss it as a clever party trick. Neither is honest about the moderate, ambiguous reality. The email I received was, in a technical sense, a perfectly serviceable piece of writing. It used my brother's name. It said the right kinds of things at the right kinds of moments. It was the kind of email I might have written if I had been writing it on behalf of someone else, on a project, in an office, with nothing personal at stake. I read it three times. And I knew I could not send it.
The reason I could not send it is more interesting than the obvious one. It was not that the email was bad. It was that the email was written in a register my brother would not recognise as mine. He has known me for thirty-eight years. He has been listening to me speak for nearly four decades, and the email I had been handed was the speech of a polite stranger who had been briefed on the case. He would not exactly know. He would feel something. The email was good and it was not mine, and the space between those two facts was where the entire situation lived.
I started writing my own email, almost without thinking about it. The AI's words were in the back of my head but I did not, with the exception of the opening sentence, use them. I kept the opening — 'There is something I have been wanting to say for a few months now' — because it sounded like the kind of sentence I might actually have produced if my own opening machinery had been working. The rest of the email took about twenty minutes. By the end, I was crying a little. I sent the email just before midnight. My brother called me the next morning. We are, as people sometimes say, okay now.
What I have been trying to work out, in the weeks since, is what the AI actually did. It did not 'write the email for me' — that part is easy to dismiss. But it also did not do nothing. What it did, more precisely, was something like this. I had been losing to the empty page for the better part of an hour. The empty page is a particular kind of obstacle, and I had not previously had a clear sense of how much of the difficulty of writing was being produced by the empty page itself rather than by the absence of things to say. The AI did not give me things to say. It gave me a page that was no longer empty. Once there was prose on the page — even prose I disagreed with — I could see what was wrong with it. The thinking I had been unable to do in the empty space happened, very quickly, in the space of reaction.
I would like to say two things, because I think they are both true and I am suspicious of essays that say only one. The first is that the experience was useful. I sent an email I had needed to send for three months. The tool had a real and clear function. The second is that I am not entirely comfortable with what the experience tells me about my own writing. Some of the friction I had been treating as the necessary work of writing turns out, in fact, to have been mechanical — the difficulty of starting, separately from the difficulty of saying. The mechanical part can apparently be removed by a tool. What worries me, slightly, is that I do not know which other parts of the writing process are similarly mechanical. I had thought my writing was more uniquely mine than this.
I would also like to flag that the email I sent had a small piece of someone else's writing inside it. My brother does not know this. It does not feel like cheating in the way handing in someone else's homework would be cheating. But it does not feel like nothing, either. I sometimes wonder whether the email I would eventually have written, in a week or a month, on my own, would have been different — and whether the difference would have been important. I cannot answer the question. I am not sure I am supposed to be able to.
What I will say, having now used the same tool a few times more, is that my use of it has settled into a particular shape. I do not ask it to write things for me. I ask it to produce a draft that is wrong in interesting ways, so that I can react to it. The pattern is becoming familiar. It is not a miraculous one, and it is not nothing. The honest reflection on a tool, I have come to think, is not whether it is good or bad in the abstract, but how exactly it is changing the shape of what you do, and whether you are still happy with the shape that is emerging.
Key Vocabulary
caveat noun
a warning or qualification before a main statement
"With the small caveat that the phrase has a worn quality."
worn (quality) adjective
(figurative) used so often it has lost its freshness
"The same slightly worn quality."
resist verb
to refuse to accept or comply with
"I have spent the past year resisting both groups."
erosion noun
(figurative) the gradual wearing down of something
"The gradual erosion of resistance."
in retrospect phrase
looking back, in the light of what happened later
"In retrospect, the argument was smaller than it had felt at the time."
register (of language) noun
the level or style of language appropriate to a context
"Written in a register my brother would not recognise as mine."
briefed adjective
given information or instructions about a topic
"A polite stranger who had been briefed on the case."
friction (in work) noun
(figurative) the resistance or difficulty in a process
"Some of the friction I had been treating as necessary."
mechanical adjective
performed by routine, without engagement of thought or feeling
"The mechanical part can be removed by a tool."
removable adjective
able to be taken away
"Similarly mechanical and similarly removable."
settle (into a shape) verb
to become regular or fixed in a pattern
"My use of it has settled into a particular shape."
in the abstract phrase
considered in general, separated from any particular case
"Whether it is good or bad in the abstract."
Questions
Comprehension
  • What does the writer say is wrong with the two standard genres of writing about AI?
    Answer
    Either the technology is declared a miracle and written about breathlessly, or it is dismissed as a clever party trick and not engaged with. Neither is honest about the moderate, ambiguous reality.
  • Why could the writer not send the email the AI produced?
    Answer
    Because it was written in a register the brother would not recognise as the writer's. The brother has been listening to the writer speak for thirty-eight years. The email read as 'the speech of a polite stranger who had been briefed on the case'. He would not exactly know — but he would feel something.
  • What does the writer say the AI actually did, in their best description?
    Answer
    It gave the writer a page that was no longer empty. Once there was prose on the page, the writer could see what was wrong with it, what they would say differently, what tone was off. The thinking the writer had been unable to do in the empty space happened in the space of reaction.
  • What two things does the writer want to say about the experience, and why both?
    Answer
    (1) The experience was useful — the writer sent an email they had needed to send for three months and they are now speaking with their brother again. (2) The writer is not entirely comfortable with what the experience tells them about their own writing — some of the friction they had treated as necessary turns out to be mechanical and removable. The writer says both because they are suspicious of essays that say only one.
  • What does the writer say is the right basis for honest reflection on a tool, near the end of the essay?
    Answer
    Not whether the tool is good or bad in the abstract, but how exactly it is changing the shape of what you do, and whether you are still happy with the shape that is emerging.
Vocabulary
  • What does 'caveat' mean, and why does the writer open with one?
    Answer
    A warning or qualification before a main statement. The writer opens with one to acknowledge that the phrase 'the first time I used AI' has become very common, and to ask the reader to set aside the expectation of breathlessness or dismissal that the phrase usually triggers. It is a small honesty up front.
  • What is the writer doing by describing the email's voice as 'the speech of a polite stranger who had been briefed on the case'?
    Answer
    Naming exactly what was wrong with the email. The metaphor compares the AI to someone — a junior staff member, a representative — who has been given the facts but has no relationship with the people involved. The image makes the failure precise: it is not bad writing, it is the wrong writer.
  • What does 'mechanical' mean here, and what is the writer's argument about it?
    Answer
    Performed by routine, without engagement of thought or feeling. The writer is suggesting that some of the difficulty of writing — particularly the difficulty of starting — turns out to be mechanical work that can be done by a tool, distinct from the deeper work of saying. The discovery is unsettling because the writer had not realised which parts were which.
  • What does 'in the abstract' mean in the closing sentence?
    Answer
    Considered in general, separated from any particular case. The writer is saying that judging a tool 'good or bad' as a whole, removed from how it is actually being used, is the wrong question. The right question is concrete and personal — about the shape it is producing in your own work.
Inference
  • Why does the writer open with the comparison to 'the first time I went on the internet' in 2002?
    Suggested interpretation
    To position the essay as deliberately aware of its own genre. The phrase 'the first time I used AI' has become routine; comparing it to a similar routine phrase from twenty years earlier admits this and asks the reader to read the essay anyway. The reasoning: the move earns trust by signalling that the writer knows what kind of essay they are writing and is trying to do it carefully.
  • What is the rhetorical effect of saying the not-starting had become a 'small ongoing fact' that the writer would notice 'on the walk to work, and at the kitchen sink, and in the soft moment between waking up and remembering what was happening that day'?
    Suggested interpretation
    It dignifies the ordinary daily presence of the unsent email. The list of small specific moments makes the avoidance physical and continuous, not a single decision. The reasoning: it justifies the writer's eventual willingness to use a new tool — they were not turning to it casually, but after weeks of small daily failure to start without it.
  • What is the writer doing by saying 'I am not sure I am supposed to be able to' answer the question of whether their solo email would have been different?
    Suggested interpretation
    Refusing the genre's pressure to resolve the ambiguity. The writer is acknowledging that the experiment cannot be re-run — the writer cannot now find out what they would have written without the tool. The reasoning: rather than producing a confident answer, the writer leaves the question open, and the open question becomes part of the essay's honesty.
  • When the writer says they 'ask it to produce a draft that is wrong in interesting ways', what is the deeper claim about how thinking works?
    Suggested interpretation
    That much of the writer's thinking happens in reaction to material rather than in generation from nothing. A draft you disagree with — even a wrong one — is more useful than an empty page, because it gives you something to push against. The reasoning: it suggests human thought is partly editorial, and that the empty page asks for a kind of work that not everyone, all the time, can produce.
  • What is the writer's underlying anxiety in saying they 'do not know which other parts of the writing process are similarly mechanical and similarly removable'?
    Suggested interpretation
    That if some of the work the writer thought of as theirs is actually mechanical, more of it might be too — and the writer cannot tell which. The reasoning: it is not the loss of any particular bit of work that worries the writer, but the loss of confidence about where their own contribution lives. The discovery is destabilising in a way that is more interesting than threatening.
Discussion
  • Is the writer right that 'the difficulty of starting' is mechanically separable from 'the difficulty of saying'? Or are the two more entangled than the essay allows?
    Discussion prompts
    Prompts for discussion: separable — many writers know the experience of having something to say but being unable to begin; the empty page produces a particular friction that is independent of having ideas. Entangled — sometimes the difficulty of starting IS the difficulty of saying; the resistance contains information about not yet knowing what one means; making it easier to start may also make it easier not to fully think. Real answer: probably both, in different cases.
  • The writer settles on a pattern of asking AI for 'a draft that is wrong in interesting ways'. Is this a sustainable approach, or does sustained exposure to AI drafts shape the writer's own future drafts in ways they will not notice?
    Discussion prompts
    Prompts for discussion: sustainable — the writer is using the tool deliberately, in a small role, and remains the author of the work. Shaping in unnoticed ways — repeated exposure to a particular style of competent prose may, over time, drift the writer's own prose toward it; the friction of being wrong-in-interesting-ways may become its own pattern. Real answer: probably both, and the writer's commitment to noticing is the only mitigation.
  • The writer keeps both 'this was useful' and 'this is unsettling' on the page at the same time. Where else in life is keeping both readings of an experience the most honest move?
    Discussion prompts
    Prompts for discussion: candidates — moving to a new country (gain and loss simultaneously), having a child, taking a new job, leaving a relationship, the death of a parent, a major piece of luck. Single-register reactions are usually false. The honest response is often two-handed. The interesting question is when keeping both becomes paralysis rather than honesty.
  • Is the writer's email — with one AI-suggested opening line — a writing of theirs, an AI-assisted writing, or a co-writing? Does the answer matter to the brother who received it?
    Discussion prompts
    Prompts for discussion: theirs — the substance is theirs, the small piece is decoration. AI-assisted — the AI did real work, even if minimal, and concealing it is dishonest. Co-writing — both contributed, though unequally. Whether it matters to the brother depends on what he would value: the apology itself, the writer's having taken time, the writer's being fully the source. Real answer: probably matters slightly, in ways that are difficult to specify, and may matter more in aggregate as more communications become assisted.
Personal
  • Is there a part of your work or thinking that you have recently realised is 'mechanical' — separable from the parts that feel uniquely yours? What was the realisation like?
    Teacher guidance
    Students' own answers. Welcome unexpected examples — formatting documents, writing meeting agendas, certain kinds of email replies, certain kinds of small talk. Common patterns: 'I realised I was treating X as personal effort when it was just a routine'. Validate the noticing. The follow-up question — what should one do with the realisation — is interesting and not easily answered.
  • Have you ever sent a message you had been delaying for a long time? What helped you send it, and what changed afterwards?
    Teacher guidance
    Students' own answers. This is a careful question. Listen first. Common patterns: 'A friend pushed me', 'It just suddenly felt possible', 'I wrote it at night', 'I sent it without re-reading'. Welcome students who say they are still waiting on a message. The writer's specific scaffolding (the AI breaking the empty page) is the model — concrete rather than mystical.
  • Is there a tool you have started using recently whose effect on your work or attention you would like to be paying closer attention to?
    Teacher guidance
    Students' own answers. Welcome answers about phones, apps, AI, social platforms, work tools, even particular routines. Common patterns: 'I check it without thinking now', 'I'm not sure I work better with it'. The exercise is to make the tool's effect visible. The writer's commitment to 'paying attention to the speed at which I am learning to work with it' is the model frame.
Writing Task
Prompt
Write a 400–600-word reflective post about the first time you used a tool, technology, or service that has since become part of your routine. The post should: (1) describe the situation concretely; (2) be honest about your initial reaction, including any resistance; (3) name the genre conventions you are working against (the technology breathless or technology dismissive essays); (4) keep both 'this was useful' and 'this is something to be careful about' on the page at the same time; (5) end without overclaiming. Use a mature, slightly self-aware voice.
Model Answer

I started using a translation app three years ago, when I moved to a country whose language I knew thirty words of, and I want to write about the first time I used it because the experience taught me something I have been trying to put down precisely ever since. There is a particular kind of essay that tends to be written about a translation app — either a celebratory one in which the writer marvels at how easy life has become, or a worried one in which the writer mourns the death of effortful language acquisition. Neither captures what actually happened. The first time I used the app was in a small bakery near my flat. I needed to ask whether the bread had nuts in it. I had a small phrase book in my coat pocket that I had used for the previous three weeks, with what I now recognise as a slightly performative diligence — circled words, careful pronunciation, the small dignity of a foreigner doing the thing properly. The phrase book had been mine in a way the app could not quite be. But on this particular morning I was tired, and the question was specifically important because of an allergy that mattered, and the colleague who had shown me the app the previous week had made it sound undramatic. I typed the question. The phone said it out loud. The woman behind the counter answered in a normal voice, the way she would have answered a customer who actually lived in the country. She told me the bread did contain nuts, recommended a different one, and I thanked her and walked home with the right loaf. I sat on a bench for a few minutes after, feeling, of all things, slightly unmoored. The transaction had been so smooth. I had been treated as someone who needed a piece of bread, rather than as a foreigner mounting a small effortful play of belonging. I had, until that morning, thought the effortful play was the point — that being visibly a beginner was somehow part of how I was learning. The app had quietly removed the visibility, and with it some of what I had been treating as the learning itself. I have used the app most days since. The local language has, in fairness, improved faster as a result, because I now have actual conversations rather than stilted ones, and conversations are how language is really learned. But I am also slightly less of a particular kind of foreigner than I was, and I am not sure where that particular foreigner has gone. The app has been useful. The app has also produced, in three years, a version of me whose relationship with the country is mediated more by software and less by visible struggle, and the difference between these two versions of me is not, I think, only a question of fluency. It is also a question of what, exactly, the previous effort had been doing.

Activities
  • Pre-reading: in pairs, students answer the intro questions and report back the most striking answer.
  • Voice analysis: students underline three sentences where the writer is being self-aware about the genre. Discuss what makes this voice work.
  • Genre awareness: in pairs, students identify the moments where the writer names the conventions of the form. Discuss what naming these conventions achieves.
  • Vocabulary in context: students take ten new words and write a single short paragraph using all ten — about a topic of their choice.
  • Discussion in groups: 'Is the difficulty of starting separable from the difficulty of saying?' Surface arguments on both sides.
  • Mock-genre rewrite: students rewrite one paragraph in the breathless 'AI is incredible' style, and another in the dismissive 'AI is nothing' style. Compare with the original.
  • Writing: students draft the 400–600-word reflective post for the writing prompt.
  • Peer review: in pairs, students read each other's drafts. Each writes one comment on what worked, one suggestion on where the voice could be stronger.
  • Cultural sharing: 'How is AI being used in my country, school, or workplace? Welcomed, debated, or restricted?' In small groups.
Duration: 50 min 🎯 Focus: Extended argumentation with concession (and yet, however), hedged generalisation, sustained metaphor (the territory, the protection), nominalisation (resistance, friction, exposure), parenthetical asides, register-shifting between practical and reflective, sustained voice across long paragraphs, deliberate awareness of and resistance to genre conventions.
Before You Read / Listen
  • Q1Is there a piece of writing about new technology you have read that did something other than the breathless or the dismissive? What was it doing instead?
  • Q2Have you noticed that some of the difficulty of a task is separable from the task itself — a kind of friction that lives in the starting rather than in the doing?
  • Q3Has a tool ever removed a difficulty you had been treating as necessary, and left you uncertain about what the difficulty had been doing?
  • Q4Is there a small daily form of self-protection in your life that you suspect is doing more work than it appears?
  • Q5What is the difference, for you, between writing that 'does the job' and writing that someone you know would recognise as yours?
The Text
I would like to write about the first time I used AI, and I am aware that this is, in 2026, a slightly tired premise. Almost every reflective essayist working in English has, at some point in the past three years, produced a piece on this subject, and the genre has settled into a small number of recognisable shapes — the breathless conversion narrative, the careful sceptic's worried inventory, the sophisticated critic's structural critique. I have read all three. I would like to try to write a fourth thing, although I am uncertain whether a fourth thing is currently available. The most I will commit to in advance is that I have a story which surprised me, and that I have spent the weeks since trying to describe what surprised me without falling into any of the available shapes.
The story is small. I needed to write a difficult email to my younger brother. We had argued at our parents' house in the summer about something that turned out, in retrospect, to have been smaller than it had felt at the time, and we had not really spoken in three months. The not-speaking had become its own quiet feature of my life — present at the kitchen sink, on the walk to work, in the moment between waking up and remembering what was happening that day. I had been sorry for some weeks. I had not been able to write the email. The longer I waited, the larger the empty space at the top of the document became.
On a Sunday evening, after an hour of producing nothing, I opened a website I had heard about but never used. I typed an honest summary of the situation and clicked the button. Two seconds later, an email appeared. What came back was a perfectly serviceable email. It used my brother's name. It opened with an acknowledgement that some time had passed. It made the apology in the appropriate place. It was also, on the third reading, unusable. The email had been written in a register that my brother — who has been listening to me speak for thirty-eight years — would not have recognised as mine. He would not exactly know. He would feel something. The email was good, and it was not mine, and the space between those two facts is the territory I would like to map.
I started, almost without making the decision, to write my own email. I kept the opening sentence the AI had produced — 'There is something I have been wanting to say for a few months now' — because it sounded close enough to the kind of sentence I might have produced if my own opening machinery had been working. The rest of the email took twenty minutes. I was crying a little by the end, in a way I had not been crying before opening the website. I sent it just before midnight. My brother called the next morning. We are, as people sometimes say, okay now.
What I have been trying to work out since is what the AI did. The two available descriptions — that it did everything, that it did nothing — are unsatisfying because they are both untrue. The accurate description is more interesting. I had been losing to the empty page for the better part of an hour. The empty page is a particular kind of obstacle, and I had not, before that evening, had a clear sense of how much of the difficulty of writing is produced by the empty page itself, separately from the difficulty of having things to say. The AI did not give me things to say. It gave me a page that was no longer empty. Once there was prose in front of me — even prose I disagreed with — the thinking I had been unable to do in the empty space happened, very quickly, in the space of reaction. The AI had supplied a draft I needed to push against, and the pushing was the writing.
This observation is, I have come to think, the most generalisable thing I have to offer. Some part of what I had been treating as the work of writing turns out to have been mechanical work that the empty page extracts from the writer before the writing proper can begin. This part can apparently be done by a tool. The deeper part — the part that involves knowing what one wants to say, and saying it in the register one's brother will recognise — cannot. The fact that the two parts can be separated is, I think, news. It suggests that the interior shape of writing, including writing of the kind that feels intimate and personal, contains layers I had not previously had a clear way of distinguishing.
I would like to flag two complications. The first is that the email I sent had a small piece of someone else's writing inside it. My brother does not know this. It does not feel like cheating in any obvious sense. The substance of the email was mine. But it does not feel like nothing either, and I am suspicious of any quick resolution of the question. I sometimes wonder whether the email I would have written on my own, in a week or a month, would have been different. I cannot run the experiment. I am sitting with the question rather than answering it.
The second complication is structural. I am writing this essay using, among other things, a brain that has been spending more time with this kind of tool over the past few weeks. I do not yet know what this exposure is doing to the shapes my own sentences want to take. The breathless and the dismissive essays in this genre tend not to address the question. The honest reflection on a tool, I have come to think, is not whether it is good or bad in the abstract, but how exactly it is changing the shape of what you do, and whether you are still happy with the shape that is emerging.
There is a final thing, which I think I have been circling without quite landing on. The email I sent took twenty minutes to write. The crying at the end was real, and was tied to something that the empty page had been holding off for three months. I am not entirely sure that the empty page had only been mechanical, in retrospect. Some of what it had been doing, perhaps, was protecting me from the feeling — letting the not-saying remain available as long as the not-starting was available too. The AI did not just produce a draft. It removed the protection. It made the writing possible, and the writing produced the feeling, and the feeling produced the email I actually sent. Whether this is a good function for a tool to perform — to dismantle, with neutral efficiency, a small daily form of self-protection — is a question I am not yet in a position to answer.
What I will say, in closing, is that I am glad I sent the email. I am glad my brother and I are speaking again. I have, since that Sunday, used the tool a few times more, in a particular settled way: I ask it to produce a draft that is wrong in interesting ways, so that I can react to it. The pattern is not a miraculous one, and it is not nothing. It is a small new thing I am learning to work with, and the speed at which I am learning to work with it is, itself, one of the things I am paying attention to. The honest answer to most questions about new technology is, in my experience, somewhere between the two available large answers, and the work is in finding the smaller, more accurate language. I do not think I have entirely found it yet. This essay is one of several attempts.
Key Vocabulary
premise noun
the basis on which an essay or argument starts
"A slightly tired premise."
inventory noun
a careful list of items or features
"The careful sceptic's worried inventory."
structural critique noun phrase
a criticism aimed at the underlying systems rather than individuals
"The sophisticated critic's structural critique."
register (of language) noun
the level or style of language appropriate to a context
"Written in a register my brother would not recognise as mine."
in retrospect phrase
looking back, in the light of what happened later
"Smaller than it had felt at the time."
machinery (of writing) noun
(figurative) the internal mechanisms that produce something
"If my own opening machinery had been working."
extract (something from someone) verb
to draw out, often through pressure or struggle
"Mechanical work that the empty page extracts from the writer."
tempo noun
the speed or rhythm of something
"Different in tone, different in tempo."
exposure (to something) noun
repeated contact with something, often shaping in effect
"What this exposure is doing to the shapes my own sentences want to take."
dismantle verb
to take apart deliberately
"To dismantle a small daily form of self-protection."
neutral efficiency noun phrase
the quality of doing something effectively without bias or feeling
"With neutral efficiency."
circle (something) verb
(figuratively) to approach without directly stating
"I have been circling without quite landing on it."
Questions
Comprehension
  • What three 'recognisable shapes' does the writer say writing about AI has settled into, and what is the writer attempting?
    Answer
    The breathless conversion narrative, the careful sceptic's worried inventory, and the sophisticated critic's structural critique. The writer is trying to write 'a fourth thing', though they are uncertain whether a fourth thing is currently available.
  • Why does the writer say they could not send the AI's email?
    Answer
    Because the email was written in a register the writer's brother — who has been listening to the writer speak for thirty-eight years — would not have recognised as the writer's. The brother would not exactly know, but he would feel something. The email was good and it was not the writer's.
  • What is the writer's most precise account of what the AI did for them?
    Answer
    It gave the writer a page that was no longer empty. Once there was prose in front of the writer — even prose they disagreed with — the thinking the writer had been unable to do in the empty space happened, very quickly, in the space of reaction. The pushing was the writing.
  • What is the writer's claim about the 'mechanical' versus 'deeper' parts of writing?
    Answer
    Some of what the writer had treated as the work of writing turns out to have been mechanical work the empty page extracts before the writing proper can begin — and this can be done by a tool. The deeper part — knowing what one wants to say, and saying it in a recognisable register — cannot. The fact that the two parts can be separated is, the writer thinks, news.
  • What is the 'second complication' the writer raises about themselves and the essay?
    Answer
    That the writer is writing the essay using a brain that has been spending more time with this kind of tool over the past few weeks, and the writer does not yet know what this exposure is doing to the shapes their own sentences want to take. The genre tends not to address this because each version is making an argument about the tool's effects on the world rather than on the writer.
  • What does the writer suggest about what the empty page may have been protecting them from?
    Answer
    That the empty page may have been protecting the writer from the feeling — letting the not-saying remain available as long as the not-starting was available too. The AI removed the protection: it made the writing possible, and the writing produced the feeling, and the feeling produced the email.
  • What is the writer's settled pattern of use after a few more uses of the tool?
    Answer
    Asking it to produce a draft that is wrong in interesting ways, so that the writer can react to it. The pattern is not miraculous, and is not nothing — it is a small new thing the writer is learning to work with.
Vocabulary
  • What does 'register' mean here, and why is it the right word for what made the AI's email unusable?
    Answer
    The level or style of language appropriate to a context. The AI's email was technically correct, but it spoke in a level of language the brother would not recognise as the writer's — too formal, too careful, too much like an office communication. 'Register' captures the precise dimension of failure: not facts, not grammar, but voice.
  • What is the writer doing by calling the empty page's work 'extraction'?
    Answer
    Treating the empty page as something that demands work from the writer — pulls effort out of them — before any actual writing can happen. The metaphor makes the difficulty external (something the page does to the writer) rather than internal (something wrong with the writer). It is a precise reframing.
  • What does 'neutral efficiency' add to the writer's description of what the AI did?
    Answer
    A precise critical edge. 'Neutral' marks the absence of judgement, care, or context; 'efficiency' marks the speed and competence. Together, the phrase captures the slightly unsettling quality of having a tool perform — without ill intent or feeling — work that had been doing something for the writer that wasn't only mechanical.
Inference
  • Why does the writer admit, in the opening paragraph, that they may not be able to write 'a fourth thing'?
    Suggested interpretation
    Because the genre is so saturated that escaping it may not be currently possible. The reasoning: by admitting the possibility of failure up front, the writer earns the right to attempt the fourth thing without claiming success. It also signals that the essay's project is honest — not promising what it cannot deliver — and aligns the writer with readers who are tired of the standard shapes.
  • What is the rhetorical effect of placing 'the thinking I had been unable to do in the empty space happened, very quickly, in the space of reaction'?
    Suggested interpretation
    It introduces a model of thought as editorial rather than generative. The writer is suggesting that human thinking, in some moods, requires material to push against — that the empty page asks for a kind of work not everyone, all the time, can produce. The reasoning: the line reframes the AI's contribution as supplying friction rather than supplying content, which is a more interesting and less defensive description than either available standard one.
  • What is the writer doing by carefully naming the 'small piece of someone else's writing' inside the email?
    Suggested interpretation
    Refusing the easy resolution. Most essays in this genre would either dismiss the inclusion as trivial or treat it as serious cheating; the writer does neither. The reasoning: by holding the small uncomfortable fact on the page without resolving it, the writer demonstrates the kind of unfinished honesty the essay is arguing for elsewhere. The shape of the discomfort is the content.
  • Why does the writer spend a full paragraph on the suggestion that the empty page was 'protecting' them from the feeling?
    Suggested interpretation
    Because it complicates the central argument of the essay. Until that paragraph, the empty page has been described as mere mechanical friction — something the AI usefully removed. The protection paragraph reopens the question: what if the friction was doing something the writer needed? The reasoning: the move stops the essay from being a tidy 'I figured out what was useful' piece, and turns it into an open question about what tools can dismantle without our noticing.
  • What is the writer doing in the closing line, 'This essay is one of several attempts'?
    Suggested interpretation
    Refusing the genre's standard closing — the small wisdom delivered at the end. The writer marks the essay as a try, not a result. The reasoning: it keeps faith with the opening admission that a fourth shape may not be currently available, and it asks the reader to read the essay as part of an ongoing process rather than a finished position. The closing modesty is consistent with the essay's overall argument about what honest reflection looks like.
Discussion
  • Is the writer's distinction between 'mechanical' work that the empty page extracts and 'deeper' work that involves knowing what to say genuinely separable, or are they more entangled than the essay allows?
    Discussion prompts
    Prompts for discussion: separable — the writer's specific experience (writing twenty minutes after the AI broke the empty page) is real and replicable; many writers know the experience of having something to say but being unable to begin. Entangled — sometimes the difficulty of starting IS the difficulty of finding what one means; the resistance contains information; making the start easy may also make easier the writing of things one has not yet thought through. Real answer: probably both, in different cases. The interesting question is when the friction was carrying information.
  • The writer suggests the empty page may have been a small daily form of self-protection. Where else in life are small frictions doing protective work that we do not notice until something removes them?
    Discussion prompts
    Prompts for discussion: candidates — having to walk to a shop (limits impulse purchases); having to physically write a letter (slows hasty replies); having to remember rather than search (consolidates memory); having to wait until tomorrow before sending (cools tempers); having to phone rather than text (forces presence); having to leave the house to see someone (gives the meeting weight). Many digital tools have removed frictions that were doing work, often without our noticing. The interesting question is which we are willing to lose and which we want back.
  • Is the writer's project — to write a 'fourth thing' that escapes the breathless, sceptical, and structural-critique shapes — possible? Or does the new shape inevitably calcify into a fourth genre that the next essayist will need to escape?
    Discussion prompts
    Prompts for discussion: possible — careful self-aware writing has always existed alongside the standard shapes, and a new technology produces new genres but also new careful writers. Inevitably calcifies — even this essay's moves (genre-naming, careful concession, refusal of summary) are themselves now recognisable; there is no escape from genre, only differential awareness of it. Real answer: probably both. The fourth thing is real and is also already becoming legible as a fourth shape.
  • If the AI's role is to supply 'a draft that is wrong in interesting ways, so that you can react to it', is the writer in a different position from a workshop participant whose drafts get critiqued by colleagues?
    Discussion prompts
    Prompts for discussion: similar — both involve generating material to react to, both involve the social-cognitive work of refining one's view by responding to other versions. Different — workshop participants are humans with stakes, judgement, and accountability; AI drafts have no author and no stake; the social texture of being criticised is part of what teaches one to be better. Real answer: similar in mechanism, different in social meaning. The writer's settled pattern uses AI for one of the things human readers also provide, but not all of them.
Personal
  • Have you ever had a friction you treated as necessary turn out, when it was removed, to have been doing something protective for you? What was it?
    Teacher guidance
    Students' own answers. Welcome unexpected examples — having to wait before sending a difficult message, having to phone rather than text, the difficulty of meeting someone in person, the slowness of a particular task. Common patterns: 'I didn't realise I needed the slowness until it was gone'. The writer's specific example (the empty page protecting from feeling) is the model — concrete and slightly counter-intuitive.
  • Is there a tool you have started using whose effect on the shape of your work or thought you would like to be paying closer attention to?
    Teacher guidance
    Students' own answers. Welcome answers about phones, search engines, AI, social platforms, calendars, notes apps. Common patterns: 'I check it without thinking now', 'I'm not sure I write the way I used to'. The writer's commitment — paying attention to the speed at which one is learning to work with a tool — is the model. Encourage specifics.
  • Is there a difficult message you have been delaying that you have not yet found a way to start? What do you think the delay is doing?
    Teacher guidance
    Students' own answers. This is a careful question. Listen first. Welcome students who don't want to share. Common patterns: 'I am not ready to feel what writing it would make me feel', 'I am not sure what I want from the conversation', 'I am avoiding hearing back'. The writer's suggestion — that the delay may itself be doing protective work — is a frame students may find useful.
Writing Task
Prompt
Write a 600–800-word reflective essay about a tool, technology, or service whose effects on you you have been trying to describe carefully. The essay should: (1) describe a specific use with concrete sensory detail; (2) name and try to write against the genre conventions of the new-technology essay; (3) acknowledge at least one complication that the genre usually skips; (4) include at least one moment where you concede that the friction the tool removed may have been doing something useful that wasn't only mechanical; (5) end without delivering a settled position. Use a mature, careful voice that allows itself to slow down and self-correct.
Model Answer

I started using a translation app three years ago, when I moved to a country whose language I knew thirty words of, and I want to write about the first time I used it because the experience taught me something I have been trying to put down precisely ever since. There is a particular kind of essay tended to be written about a translation app — either a celebratory one in which the writer marvels at how easy life has become, or a worried one in which the writer mourns the death of effortful language acquisition. Neither captures what actually happened. The first time I used the app was in a small bakery near my flat. I needed to ask whether the bread had nuts in it. I had a small phrase book in my coat pocket that I had used for the previous three weeks, with what I now recognise as a slightly performative diligence — circled words, careful pronunciation, the small dignity of a foreigner doing the thing properly. The phrase book had been mine in a way the app could not quite be. But on this particular morning I was tired, and the question was specifically important because of an allergy that mattered, and the colleague who had shown me the app the previous week had made it sound undramatic. I typed the question. The phone said it out loud. The woman behind the counter answered in a normal voice, the way she would have answered a customer who actually lived in the country. She told me the bread did contain nuts, recommended a different one, and I thanked her and walked home with the right loaf. I sat on a bench for a few minutes after, feeling, of all things, slightly unmoored. The transaction had been so smooth. I had been treated as someone who needed a piece of bread, rather than as a foreigner mounting a small effortful play of belonging. I had, until that morning, thought the effortful play was the point — that being visibly a beginner was somehow part of how I was learning. The app had quietly removed the visibility, and with it some of what I had been treating as the learning itself. I have used the app most days since. The local language has, in fairness, improved faster as a result, because I now have actual conversations rather than stilted ones, and conversations are how language is really learned. But I am also slightly less of a particular kind of foreigner than I was, and I am not sure where that particular foreigner has gone. I should pause here, because the friction the app removed may have been doing more than I have so far acknowledged. The visible-beginner posture was uncomfortable; it was also a small daily reminder that I was somewhere new, that I had work to do, that I was not the same person I had been the previous month. The discomfort was carrying information. The app made the discomfort optional, and I have, on most days since, opted out of it. The local language has improved; the daily reminder of being elsewhere has dimmed. I am not certain these are the same trade. I would like to think I am paying attention to the difference. The app has been useful. The app has also produced, in three years, a version of me whose relationship with the country is mediated more by software and less by visible struggle, and the difference between these two versions of me is not, I think, only a question of fluency. It is also a question of what, exactly, the previous effort had been doing. I do not yet have a clean answer. I am, like most users of any tool that makes a difficulty smaller, in the middle of finding out what the difficulty had been doing.

Activities
  • Pre-reading: in pairs, students answer the intro questions and report back the most striking response.
  • Voice analysis: students underline three sentences where the writer's particular voice (careful, hedging, refusing easy resolution) is at its strongest. Discuss what makes each work.
  • Genre awareness: in pairs, students identify the moments where the writer names the conventions of the form. Discuss what naming achieves and where it might overreach.
  • Vocabulary in context: students take eight new words and use each in a single sentence about a topic outside technology — politics, work, family, art.
  • Discussion in groups of four: 'The friction a tool removes is sometimes doing more than friction.' Each student takes one side for ninety seconds before swapping.
  • Translation challenge: students rewrite one paragraph in the breathless 'AI is incredible' style, and another in the dismissive 'AI is nothing' style. Compare with the original.
  • Cultural sharing: 'How is AI being used in my country, school, or workplace? Welcomed, debated, or restricted?' In small groups.
  • The friction exercise: students identify one small daily friction they have not yet examined, and write 100 words on what it might be doing for them — useful, protective, or both.
  • Writing: students draft the 600–800-word reflective essay for the writing prompt.
  • Peer review: in pairs, students read each other's essays. Each writes one comment on what worked, one suggestion on where the genre-awareness could be more earned, and one question about the tool itself.
Duration: 55 min 🎯 Focus: Periodic and balanced sentences, philosophical register, sustained metaphor (the territory, the protection, the scaffold), ironic self-awareness about the form, hedged generalisation, parenthesis as argumentative tool, holding multiple positions in tension, deliberate refusal of easy resolution, structural critique enacted within the form being critiqued, meta-essayistic awareness of the genre.
Before You Read / Listen
  • Q1Is there a piece of writing about new technology that you have read that successfully escaped its own genre? What did it do differently?
  • Q2Have you ever realised, after the fact, that a difficulty you had been treating as inability was actually doing something more like care or protection?
  • Q3Is the meta-essay — the essay aware of itself as an essay — still a useful form, or has it become as predictable as the genres it sets out to escape?
  • Q4Have you noticed your own prose or speech being shaped by repeated exposure to a particular kind of writing or speech? When?
  • Q5Is it possible to use a new tool deliberately for years without quietly changing what you take 'good' work to look like? What would deliberate use require?
The Text
There is a particular embarrassment in writing, in 2026, an essay called 'The First Time I Used AI'. The phrase has, by now, been used so many times in so many publications that even the seasoned essayist must, on opening such a piece, suppress an involuntary small wince — the same wince that greeted, in their respective times, 'I quit Facebook' and 'I went on a silent retreat'. The genre is established enough that one can predict, within a few sentences, which of the four standard shapes the essay will take. There is the breathless conversion narrative, in which the writer's relationship to the tool has the structure of a love affair. There is the careful sceptic's worried inventory. There is the sophisticated critic's structural critique, in which the tool serves mostly as an occasion for arguments the writer was already inclined to make. And there is the meta-essay, in which the writer is too aware of the previous three to commit to any of them, and produces instead a piece of writing whose subject is the impossibility of writing on the topic without falling into the available genres.
I have noticed that the essay you are now reading appears to be the fourth kind. I would like to register that I am aware of this, and to flag that the awareness does not, by itself, exempt me from the genre's pressures. Self-aware essays in this register have themselves become familiar enough that the self-awareness has begun to function as another shape rather than as an escape from the shapes. I am inside the genre, in some configuration of it, and the most I can do is to try to be honest about which configuration and at whose expense.
What I have, against this background, is a small story which surprised me. I needed to write a difficult email to my younger brother. We had argued at our parents' house in the summer about something that turned out, in retrospect, to have been smaller than it had felt at the time, and we had not really spoken in three months. The not-speaking had become its own quiet feature of my life — present at the kitchen sink, on the walk to work, in the soft moment between waking up and remembering what was happening that day. I had been sorry for some weeks. The longer I waited, the larger the empty space at the top of the document became.
On a Sunday evening, after an hour of producing nothing, I opened a website I had heard about but never used. I typed an honest summary of the situation and clicked the button. Two seconds later, an email appeared. It was a perfectly serviceable piece of writing — used my brother's name, opened with an acknowledgement that some time had passed, made the apology in the appropriate place. It was also unusable in this case. The reason is the entire content of this essay.
The email's failure was a failure of register. My brother has known me for thirty-eight years, has been listening to me speak for nearly four decades, and although he is not a careful reader of my prose on a sentence level, he has a deeply trained, mostly unconscious sense of what my voice sounds like. The email I had been handed was the speech of a polite stranger who had been briefed on the case. He would not exactly know — he would not be able to point to a particular sentence and say 'this is not you' — but he would feel something. He would feel, perhaps, that the email had been written for him rather than to him, in the same way that a corporate condolence card is written for the bereaved rather than to them. The distinction is small. The distinction is also the entire texture of intimacy, and tools that operate at scale are by their nature insensitive to it.
I started, almost without making the decision, to write my own email. I kept the opening sentence the AI had produced — 'There is something I have been wanting to say for a few months now' — because it sounded close enough to the kind of sentence I might have produced if my own opening machinery had been working. The rest of the email took twenty minutes. By the end I was crying a little. I sent it just before midnight. My brother called the next morning. We are, as people sometimes say, okay now.
What I have been trying to do since is to describe with precision what the tool actually did. The breathless answer is that it 'wrote the email for me' or 'helped me reconcile with my brother', neither of which is true. The dismissive answer is that it 'just gave me some boilerplate that I rewrote', which is true in the same way that saying a friend who got you to start a hard project 'just told you to do it' is true. Both miss the actual mechanism. I had been losing for an hour to the empty page, which is a separable kind of obstacle from the obstacle of not knowing what to say. The AI produced a draft I disagreed with. Disagreement is, in this case, a more powerful organising force than blank attention, because disagreement gives the writer a definite shape to react to. I disagreed my way to my own draft. The pushing was the writing.
This observation, if it generalises, suggests that the interior architecture of writing, including writing that feels intimate and personal, contains layers I had not previously had a clear way of distinguishing. The mechanical layer — the layer of starting, of having a draft to react to — is separable from the layer of having things to say, and the mechanical layer can apparently be performed by a tool. Writers have, for a very long time, used various scaffolds — outlines, model letters, the formal openings of business correspondence — to get past the mechanical layer. The new tool is, in one description, simply a much more flexible scaffold. In another, something different in kind. I am not yet sure which is the more accurate description, and I would like to keep being uncertain for longer than the genre usually permits.
I would like to flag two complications. The first is that the email I sent had a small piece of someone else's writing inside it. My brother does not know this. It does not feel like cheating in any obvious sense. But there is a kind of intimacy in personal correspondence whose precondition is that what you read is what your correspondent produced, and the inclusion of even a single AI-generated sentence in an email of apology to one's brother is, in a small way, a violation of that precondition. The violation is small. It is also, at scale, the texture of the next several decades of communication. There is no clean answer. The best I can do is keep the question on the page.
The second complication is structural. I am writing this essay using, among other things, a brain that has been spending more time with this kind of tool over the past several weeks than it had spent in any previous several weeks of my life. I do not know what this exposure is doing to the shapes my own sentences want to take. There are passages in this essay where I can hear, faintly, the rhythms of the kind of prose the tool produces — the careful balance, the parenthetical hedge, the sentence that summarises before pivoting. I am, in some way I cannot fully measure, learning to write in a register that is partly mine and partly something I have been reading and reacting to repeatedly. The honest reflection on a tool is not whether it is good or bad in the abstract, but how exactly it is changing the shape of what you do, and whether you are still happy with the shape that is emerging.
I should pause before going further, because I am about to attempt the move that this kind of essay has come to reward, and I want to mark it before performing it. The move is the one in which the writer, having raised serious complications, gestures toward a small reflective phrase that allows the reader to close the page reassured. The gesture is usually some version of 'and yet, the human element will always remain'. These gestures are not always wrong. They are, however, the genre's preferred way of converting honest disquiet into a comfortable closure, and I would rather, in this particular essay, not perform the conversion if I can avoid it.
What I will offer instead is a smaller and less reassuring observation. The empty page that I had been losing to for an hour was, I now suspect, not only mechanical friction. It was also doing something more like care. The empty page had been holding off the feeling that arrived, twenty minutes into writing my own draft, with sufficient force to produce tears I had not been crying before. Some of what I had been treating as inability had been, perhaps, protection — a small daily mechanism by which I had been able to remain not-yet-finished with a difficult thing. The AI did me a favour. The AI also took something I had not, until that moment, recognised that I had. The two facts coexist. The genre's pressure is to choose one of them. I would rather not.
I have, since that Sunday, used the same tool a few times more, in a particular settled pattern. I do not ask it to write things for me. I ask it to produce a draft that is wrong in interesting ways, so that I can disagree with it into my own. The pattern is not a miraculous one and it is not nothing. The pattern is, more importantly, my pattern, which is to say it is one I will need to revisit periodically, because the tool will change, my use of it will change, and the shape of my work will change with both. The honest reflection on a tool, in 2026, has to be reissued, in some form, every few months.
I am at the end of what I can usefully say in one essay. What I will say is that the email is sent, my brother and I are speaking again, the tool sits in a tab on my browser that I open with a particular small wince I am beginning to understand, and the essay is one of several I will need to write before this is anything like finished. There is light outside. I have, like most evenings I write at length, written more than I should have. I am stopping here.
Key Vocabulary
involuntary adjective
happening without conscious decision
"Suppress an involuntary small wince."
configuration noun
a particular arrangement or version
"I am inside the genre, in some configuration of it."
transactional adjective
involving an exchange, often without personal warmth
"Apologies written for institutional or transactional purposes."
boilerplate noun
standard text used repeatedly with little variation
"Just gave me some boilerplate text that I rewrote."
mechanism noun
the way something works internally
"The actual mechanism is this."
scaffold noun
(figuratively) a temporary support that helps build something
"The new tool is simply a much more flexible scaffold."
precondition noun
a condition that must be true before something else
"Whose precondition is that what you read is what your correspondent produced."
audit verb
to examine or check carefully, often officially
"Their message is being audited rather than received."
pivot (in writing) verb
(figuratively) to turn from one direction to another
"Before pivoting."
disquiet noun
a feeling of unease or anxiety
"Converting honest disquiet into a comfortable closure."
premature closure noun phrase
ending or settling something before the right moment
"A kind of premature closure."
reissue verb
to publish or release again
"Has to be reissued, in some form, every few months."
Questions
Comprehension
  • What four 'standard shapes' does the writer say AI essays settle into, and what does the writer notice about their own essay?
    Answer
    (1) The breathless conversion narrative (the tool as love affair). (2) The careful sceptic's worried inventory. (3) The sophisticated critic's structural critique. (4) The meta-essay aware of the previous three. The writer notes that their essay appears to be the fourth kind, and that being aware of this does not exempt them from the genre — self-aware essays have themselves become a familiar shape rather than an escape from the shapes.
  • What does the writer say is wrong with the email the AI produced — what kind of failure is it?
    Answer
    A failure of register. The brother has been listening to the writer speak for thirty-eight years and has a deeply trained sense of what the writer's voice sounds like. The AI's email was 'the speech of a polite stranger who had been briefed on the case'. The brother would feel that the email had been written 'for him rather than to him', the way a corporate condolence card is written for the bereaved rather than to them.
  • What is the writer's most precise account of what the AI did?
    Answer
    The AI produced a draft the writer disagreed with. Disagreement is, in this case, a more powerful organising force than blank attention, because it gives the writer a definite shape to react to. The thinking the writer had been unable to do in the empty space happened in the space of reaction — the writer 'disagreed my way to my own draft'. The pushing was the writing.
  • What does the writer say about disclosure of AI use in personal messages?
    Answer
    The writer is inclined to disclose, in future, when a sentence has come from elsewhere — but is aware that disclosure has its own costs, including making the recipient feel that their message is being audited rather than received. There is no clean answer. The best the writer can do is keep the question on the page.
  • What 'particular move' does the writer pause to mark before refusing to perform it, and why?
    Answer
    The move in which a writer, having raised serious complications, gestures toward a small reflective phrase that allows the reader to close the page reassured ('and yet, the human element will always remain', or 'what we are really learning is something about ourselves'). The writer marks it as the genre's preferred way of converting honest disquiet into comfortable closure, and refuses to perform the conversion.
Vocabulary
  • What is the writer doing by using the word 'involuntary' for the wince that greets the phrase 'The First Time I Used AI'?
    Answer
    Locating the reader's reaction in the body rather than in deliberate judgement. By calling the wince involuntary, the writer admits the genre fatigue is not a stance the reader has taken on but something happening to them automatically. It also implicates the writer — they wince too. The vocabulary choice does cooperative work between writer and reader.
  • Why does the writer use 'configuration' for their own position inside the genre?
    Answer
    Because it admits that the genre cannot be entirely escaped, only inhabited differently. 'Configuration' suggests a particular setting of variables within a fixed system, rather than freedom from the system. The word marks the writer's modesty about how much escape from the genre is actually being achieved.
  • What does the comparison to a 'corporate condolence card' do for the writer's argument about register?
    Answer
    Names the kind of failure precisely. A corporate condolence card is technically correct, demonstrates competent grammar, and addresses the bereaved by name — and is, for these reasons, especially obviously not addressed to them as a person. The image makes the AI email's failure a familiar kind, rather than an exotic technological one. It dignifies the analysis by anchoring it in something readers already know.
  • What is 'premature closure', and why is it the writer's term of choice for what the genre rewards?
    Answer
    Ending or settling something before the right moment. The writer is suggesting that the genre wants its essays to deliver settled positions, but the situation — a tool changing fast, the writer's use changing too, the shape of work in flux — does not yet justify settlement. 'Premature closure' is precise: not closure in itself, but closure happening before its time.
  • What does the writer mean by 'reissue' in 'has to be reissued, in some form, every few months'?
    Answer
    Published or released again — the way a book or document is. The writer is treating reflection on a fast-changing tool as the kind of thing one cannot do once and consider done; it has to be redone periodically as conditions shift. The word borrows the formal register of publishing to make the modesty of the claim feel earned rather than throwaway.
Inference
  • Why does the writer name and describe all four of the genre's shapes in the opening paragraph, then explicitly identify their own essay as the fourth kind?
    Suggested interpretation
    To inoculate the essay against the reader's recognition. By naming the shapes, including the meta-essay, the writer denies themselves the comfort of being safely above the genre. The reasoning: the move admits the writer is inside a recognisable form, and asks the reader to read the essay anyway, on the basis that something useful might still happen within the constraints. It is a rare kind of honesty — admitting one's own essay is not a clean escape.
  • What is the rhetorical effect of describing the AI's email as 'the speech of a polite stranger who had been briefed on the case'?
    Suggested interpretation
    It makes the failure social rather than technical. A polite briefed stranger is a recognisable type — a hospital administrator delivering bad news, a customer-service representative reading a script, a professional condoler. The image converts the AI from a piece of software into a familiar social figure whose limits are obvious. The reasoning: the essay is more persuasive about register because the failure is recognisable from contexts the reader has already encountered.
  • Why does the writer describe their own writing process — the rhythms they can hear in their own prose — as part of the essay?
    Suggested interpretation
    To put the writer's own changing shape on the page rather than treat the essay as a neutral analysis from outside. The reasoning: it acknowledges that the writer is not separate from the question they are examining; they are themselves an instance of the tool's effects. The honesty about this is the essay's most personal disclosure, and it costs the writer some of the authority the genre usually grants.
  • What is the writer doing by saying 'the tool did me a favour. The tool also took something I had not, until that moment, recognised I had. The two facts coexist'?
    Suggested interpretation
    Refusing the genre's central temptation: choosing between gratitude and grievance. The writer holds both, side by side, and resists the genre's pressure to resolve. The reasoning: the move implicitly argues that the right response to many new tools is dual rather than single, and that a sufficiently honest essay will leave the duality visible rather than collapse it.
  • Why does the writer's closing image describe the tool as sitting 'in a tab on my browser that I open with a particular small wince I am beginning to understand'?
    Suggested interpretation
    It is a precise non-conclusion. The wince is real, the use continues, the understanding is partial and ongoing. The reasoning: the image does what the rest of the essay has argued for — refusing to settle the question, while continuing to do the work. It also calls back to the involuntary wince the reader was said to share in the opening, closing the essay on a small loop that does not pretend to have moved beyond where it began.
  • What does the writer mean by saying 'the honest reflection on a tool, in 2026, has to be reissued every few months'?
    Suggested interpretation
    That essays of this kind are not finishable in a single sitting. The tool changes, the writer's use of it changes, the cultural context changes — so what was honest six months ago may not be honest now. The reasoning: it justifies the modesty of the closing without abandoning the essay's claims. It is a structural insight, not a stylistic flourish.
Discussion
  • Is the writer's distinction between the 'mechanical layer' (which the tool can do) and the 'deeper layer' (which it cannot) sustainable, or will the boundary keep moving?
    Discussion prompts
    Prompts for discussion: sustainable — there are aspects of writing that involve relationship, embodied memory, and stake that no general tool can supply. The boundary will keep moving — what was deep last year is now mechanical; what was mechanical is now obsolete; the line is historical, not essential. Real answer: probably both. Some boundary work is permanent (writing to people who have known you for decades will always exceed any tool's reach in particular ways); other boundary work will dissolve faster than we expect.
  • The writer suggests their email's small AI-written component is 'a violation' of an 'intimacy precondition'. How should this be handled at scale, when most communications come to involve some AI assistance?
    Discussion prompts
    Prompts for discussion: candidates — universal disclosure (every assisted message labelled), local norms (each relationship sets its own expectations), opacity by default (no disclosure unless asked), context-sensitive disclosure (more on personal messages, less on professional ones). Each has costs: universal disclosure may make every message feel audited; local norms create coordination problems; opacity erodes trust gradually; context-sensitive lines are unstable. Real answer: the norms are still being negotiated; the writer's open question is one many users will face.
  • Is there a useful distinction between 'a tool that helps me do what I was already doing' and 'a tool that quietly changes what I take good work to look like'? Where does the line sit?
    Discussion prompts
    Prompts for discussion: useful distinction — calculators help us do arithmetic without changing what arithmetic looks like; word processors changed what writing looks like; AI may be doing the second more than the first. Hard cases — tools rarely stay neutral over years of use; even calculators changed mathematical pedagogy. Real answer: the distinction is real but rarely clean. The writer's commitment to noticing the shape that is emerging is the responsible posture, but it is hard to maintain over time.
  • The writer refuses the closing reassurance the genre rewards. Is this refusal honest, or has 'refusing the closing reassurance' itself become a sophisticated form of closing reassurance?
    Discussion prompts
    Prompts for discussion: honest — naming the move and refusing it is the most a writer can do, and the refusal is real. Sophisticated reassurance — the meta-move has become recognisable enough that performing it now reassures readers in a particular way (this writer is sophisticated; we are in good hands; we can trust them with the discomfort). Real answer: probably both. The writer's own anxiety about this — visible in the line about the meta-essay being 'another shape' — is the most honest thing left.
  • What is the strongest critique of this essay you can construct? Be ungenerous on purpose for two minutes.
    Discussion prompts
    Prompts for discussion: that the essay is exactly what it claims to interrogate — a sophisticated meta-essay that benefits from the very genre awareness it laments; that 'I am inside the genre' is the most stylish version of the genre's standard humble performance; that the careful concession (the AI took something) is itself a calibrated move that earns the writer credibility while preserving the writer's continued use of the tool; that the email content (one apology to a brother) is too small a stake to support the essay's grand reflections; that calling other people's essays the 'breathless' or 'dismissive' shapes is a safe target chosen because those writers cannot fight back; that 'I am stopping here' is a stylish closure rather than an honest one. Then: which of these the writer would partly accept, and which they would resist.
Personal
  • Have you ever felt that a tool made you suddenly able to do something you had been treating as difficult, and then later wondered whether the difficulty had been doing something useful?
    Teacher guidance
    Students' own answers. Welcome unexpected examples — phones removing the difficulty of phoning ahead, search removing the difficulty of remembering, maps removing the difficulty of getting lost. Common patterns: 'I didn't realise I was learning by struggling until I stopped struggling'. The writer's specific example (the empty page protecting from feeling) is the model — concrete and slightly counter-intuitive.
  • Have you noticed your own writing or speech being shaped by something you have been reading or hearing repeatedly? What started shifting, and how did you notice?
    Teacher guidance
    Students' own answers. This is a careful question. Welcome answers about reading particular authors, watching particular shows, working with particular colleagues, being on particular platforms. Common patterns: 'I started using a phrase my colleague always uses', 'My emails have started sounding like the company tone'. Validate the noticing. The writer's own admission about hearing AI rhythms in their own prose is the model.
  • Is there a difficult communication you have been delaying that you suspect the delay is doing something for you — protection, or processing, or something else? You do not have to share what; only that there is one.
    Teacher guidance
    Students' own answers. This is a gentle, optional question. Listen first. The writer's frame — that the inability to start may be doing more than mechanical work — is what the question offers. Common patterns: 'I am not ready for the conversation that would follow', 'I am not yet sure what I want from it', 'I am keeping the option of having said it and not yet said it'. Validate without prying.
Writing Task
Prompt
Write a 700–900-word reflective essay about a tool, technology, or practice whose effects on you you have been trying to describe carefully. The essay must: (1) describe the tool's use with concrete sensory detail; (2) name and try to write against the standard shapes of the genre (in this case: breathless, dismissive, structural-critique, meta-essay); (3) include at least one moment where you concede that a friction the tool removed was doing more than mechanical work; (4) acknowledge how the tool is shaping your own work or thought, including the present essay; (5) end without the small reassuring gesture that the genre rewards. Use a mature voice that allows itself to slow down, hold contradictions, and refuse premature closure.
Model Answer

I started using a translation app three years ago, when I moved to a country whose language I knew thirty words of, and I want to write about the first time I used it because the experience taught me something I have been trying to put down precisely ever since. There is a particular kind of essay tended to be written about a translation app — either the celebratory one, in which the writer marvels at how easy life has become; the worried one, in which the writer mourns the death of effortful language acquisition; the structural one, in which the app becomes an occasion for arguments about the politics of English, surveillance capitalism, and what gets lost in machine translation; or the meta one, which I am about to fall into now. None of the four captures what actually happened. The first time I used the app was in a small bakery near my flat. I needed to ask whether the bread had nuts in it. I had a small phrase book in my coat pocket that I had used for the previous three weeks, with what I now recognise as a slightly performative diligence — circled words, careful pronunciation, the small dignity of a foreigner doing the thing properly. The phrase book had been mine in a way the app could not quite be. But on this morning the question was specifically important because of an allergy that mattered, and the colleague who had shown me the app the previous week had made it sound undramatic. I typed the question. The phone said it out loud. The woman behind the counter answered in a normal voice, the way she would have answered a customer who actually lived in the country. She told me the bread did contain nuts, recommended a different one, and I thanked her and walked home with the right loaf. I sat on a bench afterward, feeling, of all things, slightly unmoored. The transaction had been so smooth. I had been treated as someone who needed a piece of bread, rather than as a foreigner mounting a small effortful play of belonging. I had, until that morning, thought the effortful play was the point — that being visibly a beginner was somehow part of how I was learning. The app had quietly removed the visibility, and with it some of what I had been treating as the learning itself. I have used the app most days since. The local language has, in fairness, improved faster as a result, because I now have actual conversations rather than stilted ones. But I am also slightly less of a particular kind of foreigner than I was, and I am not sure where that particular foreigner has gone. I should pause here, because I am about to make the kind of move this genre rewards — the gesture toward a small reflective conclusion that allows the reader to close the page reassured. The standard versions are: 'and yet, true fluency requires struggle', or 'what we lose in friction we gain in connection'. These are not always wrong. They are, however, the genre's preferred way of making the trade feel like a settled bargain rather than what it actually is, which is an ongoing experiment whose results are still incoming. The friction the app removed had been carrying information. The visible-beginner posture was uncomfortable; it was also a small daily reminder that I was somewhere new, that I had work to do, that I was not the same person I had been the previous month. The app made the discomfort optional. I have, on most days since, opted out of it. The local language has improved; the daily reminder of being elsewhere has dimmed. I am not certain these are the same trade. I would also like to flag, because the genre tends to skip it, that I am writing this essay in a register that has been shaped, over the last three years, by the very kind of smoothed-out language the app produces. I can hear it in my own sentences when I read them back. I am a more efficient communicator than I was. I am also, in some way I am unable to fully measure, a slightly more anonymous one. Both are true. I am stopping here.

Activities
  • Pre-reading: in pairs, students answer the intro questions. They choose one to bring back to the class with the most surprising answer.
  • Form mapping: in groups, students annotate the essay's structure — opening genre survey, the four shapes, story setup, the failure of register, the actual mechanism, the scaffold question, the disclosure problem, the writer's own changing prose, the refused reassurance, the dual-truth ending. Discuss whether the shape itself enacts the argument.
  • Voice tracking: students mark every concession, hedge, parenthesis, or self-correction. Then they ask: where does the writer's confidence sit?
  • Critical writing: students write the strongest possible 200-word critique of the essay. Read aloud in pairs. Which critique would the writer most struggle to answer?
  • Vocabulary precision: students take eight new words and write a single paragraph using all of them — about a topic outside technology.
  • Discussion in groups of four: 'The friction a tool removes was doing more than friction.' Each student takes one side for ninety seconds before swapping.
  • The four-shape exercise: students take a topic of their choice (a habit, a tool, a cultural shift) and outline the breathless, dismissive, structural-critique, and meta versions of an essay about it. Then attempt a fifth thing.
  • Comparative reading: students place the C1 and C2 texts side by side and identify the moves that only the C2 text makes (the four-genre opening, the corporate-condolence-card metaphor, the disclosure problem, the dual-truth ending). Discuss whether the additional moves are earned by the topic.
  • Writing: students draft the 700–900-word reflective essay for the writing prompt.

⭐ Ratings & Comments

How useful did you find this text? Leave a rating and a comment to help other teachers.

Your rating:
No rating