Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124

You open a blank document and place your hands on the keyboard.
You know what you want to say. The subject is familiar. A few years ago, you would have started writing, even if the first paragraph was rough.
Now you stare at the cursor.
After several minutes, you open ChatGPT and type, “Write an introduction about…”
The response appears immediately. You may not love it, but at least the page is no longer empty.
This is becoming a real concern for students, professionals and writers. People who once wrote emails, reports, essays and stories independently now feel unable to begin without asking AI for an outline, opening sentence or complete first draft.
Has ChatGPT actually damaged their writing ability, or have they simply become accustomed to an easier starting point?
Frequent dependence on ChatGPT can make independent writing feel harder because the tool begins performing the mentally demanding parts of writing for you.
These include finding a starting point, recalling relevant ideas, choosing an argument, arranging information and turning unclear thoughts into sentences.
The ability has probably not disappeared permanently. It may have become less practised, while your confidence in using it has fallen.
The problem is often not that you have forgotten grammar. It is that you no longer trust yourself to create the first imperfect version.
Writing may look like typing words onto a page, but several mental tasks happen before a sentence appears.
When you write without AI, your mind must:
ChatGPT can perform many of these jobs in seconds.
That saves time, but it also changes your role. Instead of generating and organising the material, you may only review what the AI produced.
Reviewing is still a skill. It is not the same skill as creating.
If AI repeatedly performs the beginning, middle and structure of the task, your independent writing process receives less practice.
A useful way to understand the problem is to divide writing into four jobs.
You search your memory for examples, facts, experiences and language.
You decide which ideas matter and which should be removed.
You arrange those ideas so one thought leads naturally to the next.
You turn the organised thought into words that sound like you.
When you ask ChatGPT to write an entire draft, it can perform all four jobs.
You may then correct facts, adjust the tone and remove awkward phrases, but most of the original decisions have already been made.
The final document can look polished while giving you very little practice in producing it.
A blank page has always been uncomfortable.
Before generative AI, writers had to move through that discomfort. They wrote a weak sentence, deleted it, tried another opening and gradually discovered what they wanted to say.
ChatGPT provides an immediate escape.
Once your brain learns that discomfort can be removed within seconds, remaining with the blank page starts feeling unnecessary. Opening ChatGPT becomes the automatic first step.
This does not necessarily mean your brain has been damaged. It may mean your tolerance for the early uncertainty of writing has changed.
The blank page once meant, “Start exploring.”
Now it can feel like, “Why struggle when a draft is one prompt away?”
This is similar to why the brain begins resisting difficult work when an easier and faster option is always available.
Many people say they “cannot write anymore,” but that statement can hide several different experiences:
These experiences do not prove that the underlying ability is gone.
ChatGPT shows you a clean response without showing the failed attempts that human writing usually requires. Comparing your first draft with an AI-generated response is therefore unfair.
Your draft contains the search process. The AI response hides it.
If you expect your first sentence to sound finished, normal writing will begin to feel like failure.
A human first draft is often repetitive, uneven and incomplete.
That is normal. Its purpose is to give your thoughts a visible form so you can improve them.
ChatGPT changes the reference point. You become accustomed to receiving:
Your unaided draft may look weak beside it.
The risk is that you stop writing before giving your own ideas time to develop. You judge the roughness of the process as evidence that you are no longer capable.
But polished language is not always stronger writing. A smooth AI paragraph may contain vague claims, repeated structures and no personal insight.
A rough human paragraph may contain the one original observation that makes the article worth reading.
When you build a paragraph yourself, you usually remember why each part is there.
You know which example led to the conclusion. You remember rejecting one phrase and choosing another. The words are connected to a series of decisions you made.
When ChatGPT generates the draft, those decisions occur outside your own writing process.
You may understand the finished paragraph without feeling that it belongs to you. You might also find it harder to recall the argument later because you did not construct it step by step.
A widely discussed 2025 preprint, Your Brain on ChatGPT, compared people completing essay-writing tasks with an AI assistant, a search engine or no external tool. In that experiment, the AI-assisted group showed lower measured neural engagement, weaker recall of their own essays and a lower reported sense of ownership.
However, this research should be interpreted carefully. The study involved a relatively small number of participants, focused on a particular essay task and was released as a preprint rather than settled proof that ChatGPT permanently harms the brain.
Its most useful question is not whether AI “destroys intelligence.” It is whether receiving finished language changes how deeply a person engages with the writing process.
Writing style develops through thousands of small choices.
You choose whether a sentence should be short or detailed. You use certain words more naturally than others. You decide when to be direct, humorous, uncertain or emotional.
ChatGPT also has patterns. Its writing may repeatedly use:
After editing AI drafts every day, you may begin copying those patterns even when writing independently.
This does not mean AI has literally replaced your voice. It means the language you repeatedly read and approve can influence what begins to feel normal.
Some writers then experience a strange problem. AI writing does not sound like them, but their unaided writing no longer feels familiar either.
They have spent more time correcting generated language than producing their own.
Creating a detailed prompt requires thought.
You may define the audience, tone, purpose, structure and key points. That is meaningful intellectual work.
But prompting is not identical to writing the final piece.
If you tell ChatGPT:
Write a persuasive email to my manager requesting flexible hours. Keep it professional and mention my productivity, travel time and availability.
You have identified the purpose and supporting points. You have not practised building the argument, choosing its rhythm or phrasing the request yourself.
Prompting can strengthen planning while reducing practice in expression.
The distinction matters because people sometimes believe they are maintaining their writing ability simply because they are giving AI detailed instructions.
They may be directing writing rather than performing it.
Some users begin with ChatGPT because it appears faster, then spend longer fixing the result.
They repeatedly ask it to:
The first draft took seconds, but the editing loop continues for an hour.
This can leave you mentally tired without giving you the satisfaction of having written something yourself. It may contribute to the feeling of being mentally drained after using AI tools.
At some point, managing generated text can require more effort than writing a shorter, clearer version independently.
Students do not write essays only to produce documents. They write to learn how to form an argument.
The struggle to remember information, connect evidence and explain an idea is part of the educational task.
If ChatGPT performs that struggle, the submitted essay may look competent while the student remains unable to explain it without the screen.
This becomes visible during:
The issue is not simply academic honesty. It is whether the student is developing the ability the assignment was designed to practise.
This is one practical way AI may weaken critical thinking when it replaces the learning process rather than supporting it.
The problem is not limited to students.
Employees now use ChatGPT for emails, reports, proposals, meeting notes, presentations and performance reviews.
At first, AI may only help with difficult documents. Gradually, it can become the starting point for ordinary communication.
A person who once wrote a five-line email directly may begin asking ChatGPT to create it, shorten it and adjust its tone.
The task is completed, but a small act of independent expression has been outsourced.
Repeated across hundreds of messages, this can affect confidence. When AI is unavailable or unsuitable, even a simple document starts feeling more difficult than it should.
Your use may deserve attention if:
These signs do not mean you must stop using AI. They suggest that the tool is doing more of the core writing than you intended.
Probably not.
Writing is a practised skill. When it receives less use, it can feel slow and uncertain. When practised again, fluency can return.
The first attempts may be uncomfortable because your expectations have changed. You are comparing your rough process with instant generated prose.
Do not use the discomfort of returning as proof that recovery is impossible.
Your ability to notice that something has changed is itself important. It means you still recognise the difference between generating your own language and approving someone else’s.
Give yourself ten minutes to write without correcting anything.
Do not aim for a polished introduction. Start with:
What I am actually trying to say is…
This removes the pressure to sound finished.
Before asking ChatGPT for help, write:
The outline can be rough. Its purpose is to keep the structure connected to your thinking.
A balanced process can have three stages:
This keeps AI in the review process without handing it authorship of the whole piece.
Instead of saying:
Write this section for me.
Try:
Ask me five questions that will help me develop this section.
Your answers become the source material.
If you are stuck on ideas, ask AI to suggest angles but write the sentences yourself.
If you need language correction, write the full passage first and request targeted feedback afterward.
Avoid outsourcing both thinking and expression at the same time.
Choose one small form of writing:
The purpose is not to reject AI. It is to maintain direct access to your own words.
Look at work you created before using AI heavily.
Notice the expressions, humour, sentence rhythms and observations that felt natural to you. Do not copy the old style mechanically. Use it as evidence that your voice existed before the current doubt.
If ChatGPT explains something well, close the response and write the idea again from memory.
This forces you to process the meaning instead of rearranging generated sentences.
Not every message needs to sound professionally edited.
A clear human email with one awkward sentence may communicate more honestly than a perfectly polished paragraph that sounds like everyone else.
ChatGPT can still be useful for writing when its role is limited and intentional.
It can help you:
The key difference is timing.
AI used after you think can improve the work. AI used before you think may quietly replace the process.
That is why over-relying on ChatGPT is not only about how often you use it. It is about which parts of your mental work you consistently hand over.
ChatGPT probably has not erased your ability to write.
It may have changed what you expect writing to feel like.
You have become accustomed to immediate structure, polished sentences and a way around the blank page. Your own writing now feels slower because it includes searching, choosing, failing and revising.
Those are not signs that writing has stopped working.
They are writing.
The goal does not have to be writing without AI forever. It is being able to write when AI is unavailable, unnecessary or unable to express what only you know.
Before asking ChatGPT for the first sentence, write one imperfect sentence of your own.
That small act keeps the beginning of the thought with you.