What the AI Writing Survey Says About Jobs, Freelancers and Fiction Authors
TL;DR: The survey “AI and the Writing Profession” asked 1,481 writers about AI and got two forecasts for the same weather. Freelancers who use AI heavily were the most likely to report more work and more income. Nonusers were the most likely to report losses and expect worse. Fiction authors split into camps that barely overlap. The trade is splitting, and how a writer uses the tool decides which side they land on.
Ask a thousand people in one town what tomorrow’s weather will be and you expect one forecast with a little noise. Ask a thousand writers what AI is doing to their trade and you get two forecasts that don’t overlap.
One says clear skies: more work, more income, better pages. The other says the storm is already on the roof. Lost clients, friends out of work, a trade that will be smaller in five years. Both groups live in the same town and answer to the same clients and editors.
What decides the forecast is almost entirely how much the writer uses the tool.
The headline numbers from this survey have been quoted everywhere, twice on this site. The sections on jobs, freelancers and fiction authors got less attention, and they’re the ones a working writer needs.
What is the “AI and the Writing Profession” survey?
“AI and the Writing Profession” is a 2025 survey of 1,481 working writers, published by Gotham Ghostwriters with design and analysis by Josh Bernoff. It counted 1,190 writing professionals and 291 fiction authors, and responses came in during August and September 2025.
It’s a non-probability online sample, so treat it as a strong signal. If it were random, the margin of error would be about three points for professionals and six for fiction authors. The sponsor paid for the study and, per the methodology page, didn’t review the findings before publication.
The authors also say they used AI only for research on the report design and for picking which comments to quote. That’s a better disclosure than most books carry.
Many are. Of the 806 freelancers and small-agency writers surveyed, 45% said AI had reduced demand for their work and 40% said it had cut their income. The doom forecasts are built on those numbers, and they’re real.
Split them by how heavily each writer uses AI and the picture changes.
Among advanced users, the people who use AI daily for six or more kinds of tasks, 21% reported more demand and 25% reported higher income. Among nonusers, 2% reported more demand and 1% reported higher income, and 44% reported lost income, the worst of any group.
Freelancers by AI use class. Source: AI and the Writing Profession, gothamghostwriters.com/ai-writer.
A third of advanced users still lost demand, so heavy use is no shield. But the gaining side of the market belongs almost entirely to them. The same storm hit every roof. Some writers had already moved the furniture.
I’m a freelance ghostwriter, so this is my side of the market.
A client who buys blog posts by the word always had a cheaper option somewhere, and now the cheapest option ever offered sits in a browser tab. That client was leaving anyway. The ones who stay pay for judgment, interviews, structure and a voice that sounds like a person, and a writer who uses AI to speed up the rest has more hours to sell those.
How many writers have lost jobs to AI?
Fewer than the headlines suggest and more than the optimists admit.
Only 10% of writers at companies and agencies with more than ten people said their organization had laid off writers expecting AI to cover the work. Yet 43% of writing professionals know a writer who lost a job to AI, and 5% know more than ten. One layoff ripples through a whole network of friends, and a freelancer who loses a client counts as a lost job to everyone who knows her.
The surprise was who made the call. Among writers at those larger organizations, 67% decided to use AI themselves. Managers suggested it for 19% and required it for 14%.
The boss forcing a chatbot onto a reluctant staff is mostly a myth. Writers picked up the tool on their own, the way they picked up word processors.
Who decided, plus layoffs and lost jobs. Source: AI and the Writing Profession, gothamghostwriters.com/ai-writer.
The report’s authors have a blunt warning for managers. Cutting staff and expecting the survivors to make up the difference with AI is, in their words, “a prescription for disaster.” They suggest using the advanced users already on staff as mentors. I agree with both halves. A company that fires the writers who know the tool, keeps the ones who don’t and hands out logins has bought itself a flood of flat copy and invented facts.
Which writing jobs expect the worst from AI?
All of them expect worse. Overall, 73% of writing professionals expect fewer opportunities over the next five years, and 16% expect more.
Speechwriters were the gloomiest at 80%, then journalists at 77% and PR people at 76%. Book ghostwriters sat at 71%, nonfiction authors at 70% and content marketers at 68%. Technical writers were the least pessimistic, and even there 59% expected a decline. The split by AI use cuts deeper than any split by job. Among nonusers, 82% expect fewer opportunities and 9% expect more. Among advanced users, 51% expect fewer and 36% expect more.
Five-year outlook by AI use and by profession. Source: AI and the Writing Profession, gothamghostwriters.com/ai-writer.
Even the heaviest users lean pessimistic. They’re just far less sure the sky is falling, and I think I know why. Nonusers are forecasting the trade. Heavy users are forecasting the trade and their own place in it, and they’ve seen where the tool breaks.
Do fiction authors use AI?
Four in ten do. In the survey, 42% of fiction authors use AI. Another 9% rarely do, and 49% never touch it.
The users are enthusiastic. Of those, 87% said AI makes them more productive, by an average of 57%, and 60% said it improves their writing. Only 11% use it weekly to make text they publish without significant editing.
The nonusers live in a different country. Every one of them agreed that training AI on their work without permission is unfair, and every one agreed AI content should always be labelled. Just 1% called AI a positive force for the profession, none said it helped their career and 26% have considered quitting over it.
Fiction authors who use AI versus those who don't. Source: AI and the Writing Profession, gothamghostwriters.com/ai-writer.
Unanimous agreement is rare in any survey.
For that group, AI is a moral question with the answer already filled in.
People are entitled to that view, and the training question has real legal weight. What I object to is how the view gets enforced in writing groups. Someone who accuses a stranger of theft on a hunch, piles on a person who asked an honest question or decides who counts as a real writer is practicing cruelty with a cause stapled to it. I wrote about that in AI Derangement Syndrome. It’s a tool. Use it or don’t.
Should AI-assisted writing be labelled?
Labelling splits the camps harder than anything else in the report. Among nonusers, 91% said content made with the aid of AI should always be labelled. Among advanced users, 25% agreed.
Each group pictures something different. A nonuser pictures a machine writing the book. An advanced user pictures a title suggestion, a search, a word that wouldn’t come and an outline that got torn up and rebuilt. Label every page that involved a search and you label nearly everything published this year. Books raise a narrower question, and I went through it in AI disclosure and what you owe readers. Readers are owed honesty about who did the thinking, the stories and the judgment.
Ask writers about AI and you get two forecasts for the same weather. Heavy users see work growing. Nonusers see a storm. The tool doesn’t decide which writer you are. How you use it does. – Richard Lowe
Why do AI users and nonusers see the writing profession so differently?
The report scored every respondent twice, once for how much they use AI and once for how they feel about it. The two scores correlated strongly, at about 0.7, and only 13% of respondents came out net positive.
Correlation doesn’t say which way the arrow points. Some writers tried AI and grew to like it. Others disliked it from the start and never touched it. The income gap runs the same direction: advanced users reported a median of $120,100 against $73,400 for nonusers, 64% higher. Both camps share one fear. Nine in ten writers worry about hallucinated facts, and they’re right to, because a writer who uses AI still has to check every claim it produces.
My position hasn’t moved.
The tool used to write a page doesn’t matter. The use does, and so does the result: the memoir sounds like the person, the horror horrifies, the reader trusts the facts. AI working alone produces flat work at best and slop at worst. A human brain directing it as an assistant is a different animal. I use AI every day as a digital assistant that helps me write, and I still do the writing.
What should a working writer take from the AI writing survey?
Learn where the tool breaks before you decide what you think of it. The report tells objectors the same thing: they’ll argue their case better once they know its limits.
Sell what a prompt can’t produce. Interviews, structure, judgment, voice and the nerve to tell a client the chapter doesn’t work are what the freelancers on the gaining side of that chart are selling.
For many, yes. In the 2025 survey AI and the Writing Profession, 45% of freelancers said AI reduced demand for their work and 40% said it cut their income. Among the heaviest AI users, 21% reported more demand and 25% higher income.
Is AI replacing writers at work?
In the survey, 43% of writing professionals know a writer who lost a job to AI, and 5% know more than ten. Only 10% at larger organizations reported AI-driven layoffs where they work.
Are companies making writers use AI?
Mostly no. Among writers at companies and agencies with more than ten people, 67% chose to use AI themselves, 19% said managers suggested it and 14% said managers required it.
Which writers are most worried about AI?
Speechwriters, with 80% expecting fewer opportunities in five years, followed by journalists at 77% and PR professionals at 76%. Technical writers were least worried at 59%.
Do novelists use AI to write their books?
About four in ten fiction authors use AI, mostly for brainstorming, search and finding words. Only 11% use it weekly to make text they publish without significant editing, and 49% never use it.
Do writers think AI use should be disclosed?
It depends on whether they use it. 91% of nonusers said content made with the aid of AI should always be labelled, against 25% of heavy users.
Do writers who use AI make more money?
In the survey, the heaviest AI users reported a median income of $120,100 against $73,400 for nonusers, 64% higher. It can’t tell whether AI raised those incomes or high earners adopted AI first.
Richard Lowe is a professional ghostwriter and author with 113+ books authored and 54+ ghostwritten. Before writing full time he spent 33 years in enterprise technology, including 20 years as Director of Computer Operations and Technical Services at Trader Joe's. He writes nonfiction, fiction and memoir, and works with executives and experts on books that build authority.
The views and opinions expressed in this blog post are solely those of Richard Lowe and are based on personal experience and research. This content is for informational purposes only and should not be construed as professional legal, financial, accounting, or business advice. Always consult with qualified professionals before making important business or legal decisions. Richard Lowe is not a lawyer, accountant, or licensed professional advisor, and this content does not establish any professional relationship.
