For two years the AI debate among writers was all shouting and no numbers, and I was sick of it. This survey finally puts real data on the table, and some of it will make both camps uncomfortable.
The debate about AI and writing has been loud and emotional. Writers who love it call it a productivity revolution. Writers who hate it call it a plagiarism machine. Everybody has an opinion. Almost nobody had a number until now. That hasn’t slowed the arguing down at all.
Now we do.
In 2025, Gotham Ghostwriters and Josh Bernoff ran the first large-scale survey of professional writers on AI, reaching 1,481 respondents including 1,190 writing professionals and 291 fiction authors. Bernoff brought 30 years of survey analysis experience and previously led the 2024 Business Book ROI Study. The full report is available at gothamghostwriters.com/ai-writer.
I read the entire report. I use AI daily, I’ve written 46 handbooks on AI-enhanced writing, and I consult with businesses on AI implementation, and the findings confirmed what I’ve been telling clients and fellow writers for over a year. The data tells a clear story, and it’s not the one most people expect.
Are writers divided over AI?
AI can’t interview you for ten hours to find the stories that make your book convincing.Share on X
The study’s most important finding is that how writers feel about AI depends almost entirely on how much they use it. Writers who use AI heavily think it’s making them better at their jobs, while writers who don’t use it think it’s destroying the profession, and there’s almost no middle ground. For more, see why cursing at ChatGPT works and why politicians with speechwriters still need a book.
Among the most advanced AI users (the 15% who use it daily across six or more different tasks), 57% said AI is a positive force for the writing profession. Among nonusers, 3% said the same. Those are two completely different realities.
The advanced users report a median income of $120,100. Nonusers report $73,400. That’s a $47,000 gap. You can argue about causation versus correlation, but the pattern is consistent across the data. The writers who use AI the most are earning the most and feeling the best about their careers. The writers who refuse to use it are earning less and feeling threatened.
What Do Writers Use AI For?
Most uses are small jobs around the writing, and few writers publish AI text unedited.
Mostly for small jobs around the writing: suggesting titles and headings, replacing web search, brainstorming, finding the right word, and research. Very few let AI write text they publish unedited.
That’s where AI belongs. The small jobs around the writing save real time, and none of them touch the part a reader pays for. The writers who hand over the sentences themselves are the ones I’d worry about.
The public fear is that writers are using AI to generate finished text and publish it without editing, and most of the conversation about AI and writing assumes that fear is right, but the study’s data says otherwise.
Only 7% of writing professionals use AI to generate text that gets published without significant editing. Less than 1% do that daily. Most of them use it as a tool inside the process they already had. Almost nobody is letting it write the book.
The most common uses: suggesting titles and headings (72% of AI users), replacing web search (71%), brainstorming (68%), finding the right words for a concept (68%), and research (63%). These are the tasks that slow writers down between the parts of the job that require actual expertise. AI handles the mechanical friction so the writer can focus on the thinking.
Among ghostwriters (my profession), search is the dominant AI task.
The study found that search is the only AI task more than three out of four ghostwriters embrace. This makes sense. Ghostwriting is built on understanding the client. AI helps with research and organization. The strategic thinking, the interviewing, the narrative structure: that’s still the ghostwriter.
One ghostwriter quoted in the study described uploading ten or more hours of transcribed client interviews into AI to help draft outlines and match the client’s natural voice and tone. That’s my process. The AI isn’t writing the book. The AI is helping me work with the raw material from interviews more efficiently so the final product is better and faster.
How much more productive are writers who use AI?
Productivity is the number people love to wave around, and it’s the one I trust least. Finishing a draft faster means nothing if the draft is hollow, and a writer who saves a week and loses the reader has made a bad trade.
Three out of four AI-using writers said it makes them more productive. Among advanced users, 92% said so. The average productivity increase reported was 31%.
Productivity isn’t the whole story, either. Among advanced users, 59% said AI has improved the quality of their writing. Quality, on top of speed. They’re producing better work faster. Only 9% of AI users said it made their writing worse.
That matches my experience. AI doesn’t make me a better writer in the sense of better ideas or deeper insight. It makes me better by giving me back the hours I used to spend on research, organization, finding the right reference, and checking consistency across a 60,000-word manuscript, so they go to the work that needs me.
The Fear Is Also Real
I don’t dismiss the fear. Writers who spent years learning the craft are watching clients and publishers treat words as something a machine can churn out for nothing, and that pressure lands on working writers whether or not the machine’s output is any good.
The study doesn’t sugarcoat the other side. Writers who don’t use AI are afraid, and some of their fears are justified.
Forty-five percent of freelancers reported reduced demand for their work due to AI. Forty percent said AI has reduced their income. Forty-three percent of all writing professionals know someone who has lost their job to AI. One in four writers has considered quitting the profession because of AI.
Among nonusers, 89% worry that corporate leaders will replace writers with AI. And 82% of nonusers predict that opportunities for professional writers will decrease in the next five years. Even among the overall writing population, 73% expect a decline.
These fears aren’t irrational. Companies are using AI to replace writers. The work that’s disappearing first is the work that was most commoditized to begin with: generic blog posts, SEO content, basic marketing copy. Content that never required deep expertise is being automated because it can be.
But the study also shows that only 10% of writers working in corporations and agencies have seen colleagues laid off due to AI. The perception of job loss is running far ahead of the reality. The writers losing work are disproportionately those doing commodity work that was already undervalued.
What Does AI Adoption Mean for Ghostwriting?
For anyone hiring a ghostwriter, the question to ask is who writes the sentences. I use AI every day as a digital assistant for transcripts, session summaries and first-pass outlines, and I still write every book myself. A ghostwriter who can’t tell you where their own line sits hasn’t thought about it hard enough.
Sixty-eight percent of book ghostwriters in the study use AI. That’s higher than journalists (44%) but lower than thought leadership writers (84%) or PR professionals (73%). Ghostwriters are adopting AI selectively, using it where it accelerates their process while keeping the human elements (interviewing, strategic positioning, narrative craft) firmly in their own hands.
One ghostwriter in the study said something I hear constantly from peers: “I’m frustrated that potential clients think it’s a replacement for human writers.”
That’s the real challenge. The threat to ghostwriters isn’t that AI will write better books. It won’t. The threat is that clients will believe AI can write their book and skip the ghostwriter entirely. I talked about AI and the future of creative work on Tequila and Tech Talk.
They’ll learn the hard way that it can’t. AI can’t interview you for ten hours to find the stories that make your book strong. It can’t call your clients to capture transformation narratives. It can’t structure 60,000 words into a narrative that keeps a reader engaged from introduction to conclusion. And as the Gotham ROI Study showed, ghostwritten books generate four times the revenue of other books. The value is in the human expertise.
Ghostwriters who refuse to use AI will fall behind ghostwriters who use it. The income data is clear. The productivity data is clear. AI is a tool that makes good ghostwriters better and faster. Refusing to use it is a competitive disadvantage dressed up as a principled stand.
The Concerns Everyone Shares
One area where AI users and nonusers agree: the problems are real.
Ninety-one percent of all writing professionals are concerned about hallucinations. This is the biggest risk in AI-assisted writing. AI generates plausible-sounding content that’s factually wrong, and it does so with complete confidence. Any writer using AI who’s not fact-checking every claim is producing unreliable work. I’ve seen this in my own practice and I’ve written extensively about it. AI hallucination is the single biggest danger in AI-assisted writing.
Eighty-one percent are concerned about AI being trained on copyrighted content without permission. Eighty percent are concerned that AI is contributing to bland, boring writing. These are legitimate concerns that the industry needs to address.
And here’s the stat that should worry everyone who cares about quality: 80% of writing professionals think AI is feeding the trend toward blander, more boring text. The 7% who publish AI text without editing are a big part of the reason, because they’re flooding the internet with generic, voiceless content that drags down everything readers see.
The writers who use AI well (as a tool within a human-driven process) are producing better work. The writers and non-writers who use AI as a replacement for thinking are producing garbage. Both things are true at the same time, and the study captures both.
Where This Goes
My bet is that the shouting between the two camps fades, and what’s left is a split between writers who check the machine’s work and writers who don’t. Readers will sort those two groups out without any help from a survey.
The study paints a profession in transition. The writers who adapt will earn more, produce more, and find more opportunities. The writers who refuse to engage with AI will earn less and watch their market shrink. The writers who use AI as a crutch instead of a tool will produce low-quality work that damages their reputation.
The path forward is what I practice every day and what I teach in my AI consulting work: use AI for what it does well, keep your expertise firmly in control of the process, fact-check everything, and never let the tool replace the thinking. And if you’re writing a book, don’t write the manuscript with AI.
Use it for research, outlines, brainstorming, and organization. Write the book yourself or hire a ghostwriter. The data now confirms this is the strategy producing the best financial and professional outcomes for writers across the industry.
If you’re a writer figuring out how to integrate AI into your process, or a business leader trying to understand what AI means for your content strategy, start with a conversation. My AI-Enhanced Writer’s Library covers the practical side in detail across 46 handbooks.
Data referenced from “AI and the Writing Profession: A Complete Survey & Analysis” by Gotham Ghostwriters and Josh Bernoff, WOBS LLC (2025). The study surveyed 1,481 writing professionals and fiction authors.
AI and Professional Writing FAQ
The writers screaming that AI will destroy the craft and the ones promising it’ll write their books for them are both wrong. The data says what I see in my own work: it’s a tool, and it pays off for people who already know how to write.
The book on this: The Birth of the Augmented Human is 212 pages on what goes right when writers use the tools deliberately. That’s what most of that sixty-one percent are doing.
