A cohort of novelists, magazine editors, and marketers is deliberately rewriting their prose to look nothing like ChatGPT output — killing the em dash, cutting the rule of three, dropping typos back in on purpose, and leaning hard into first-person anecdote. Playboy editor-in-chief Philip Picardi opened submissions in May and got around 1,000 pitches; after experimenting with Claude himself, he says the majority had been generated with AI. Pitchfork's head of editorial content Mano Sundaresan has since sent writers a memo banning AI in review copy. The pattern is consistent enough that novelist Laura Brooke Robson calls it a literary counterculture in the making.
Robson wrote her novel Love Is an Algorithm — about the cofounder of an AI dating app and a musician navigating LLMs — while actively avoiding what she calls wishy-washy metaphors, and taking a more winding path to steer around em dashes and tripartite lists. She says the style now takes shape through negation: no passive voice, no not-x-but-y constructions, no obvious LLM tells. What replaces them is idiosyncrasy — invented phrases, odd punctuation, the specific first-person anecdote that a language model has no access to.
The stakes are not abstract. In March, Hachette pulled Mia Ballard's horror novel Shy Girl after online speculation that she had heavily used AI, an allegation she rejected. Ballard says an editor she hired ran the book through AI software without her knowledge, and she is now editing new work herself, calling authentic-feeling prose her number one priority. Steven Rosenbaum's book The Future of Truth was found to include fabricated quotations generated by AI, which Rosenbaum later admitted. The publishing industry is now sensitive enough to AI signatures that being suspected is nearly as damaging as being guilty.
Key facts
- 01Playboy editor Philip Picardi received around 1,000 pitches in May and says the majority were generated with Claude or similar LLMs.
- 02Hachette pulled Mia Ballard's horror novel Shy Girl in March after online allegations of AI use, which Ballard rejected.
- 03Pitchfork editorial head Mano Sundaresan has sent writers a memo banning AI in review copy since taking the role in July 2024.
- 04Ben Horwitz's Sinceerly app A/B-tested emails to Fortune 500 CEOs; only the misspelled, lowercase ones got opened.
- 05John Warner's book More Than Words argued last April that AI could reinstate the value of good writing rather than replace it.
Australian novelist Ellena Savage says AI has stolen syntactical conventions she once held dear, including the em dash and the rule of three. Richard Fisher, an honorary professor at University College London and editor at Aeon, says he now encourages writers to be more human and more conversational — refusing, he says, to sand out the rough edges, because AI cannot write prose as spiky or idiosyncratic as humans. Sundaresan, who took over Pitchfork editorial in July 2024 as LLMs became mainstream, pushed writers toward their own voice and first-person perspectives; in hindsight, he says the shift could have also come from this AI turn. Pitchfork is owned by Condé Nast, which also owns WIRED, the outlet that first reported the trend.
Playboy's Picardi has taken the counterstyle further into commissioning. After the May flood of AI pitches, he says the magazine committed to first-person human experience and formally adventurous writing, invoking bell hooks refusing to capitalize her name as the kind of writing AI can't replicate. Clickhole writer Jessye McGarry pushed the joke to its logical endpoint with an article titled 5 Reasons to Love Summer (NO AI), filled with intentional typos and pleas of human authorship. McGarry says leaving a spelling mistake in — beaitiful sunshine — took more work than typing it correctly, because she had to keep undoing autocorrect.
“The key thing from that experience that became really important to us at Playboy was that we needed first-person human experience.”— Philip Picardi, Editor-in-chief of Playboy
On LinkedIn, the counterstyle has hardened into a genre of its own. Ray Slater Berry, founder of marketing agency dslx, says he always pushes for first-person content because he wants to see the blood, sweat, and tears that went into it, and the agency will always be writing in the first-person from hereon out. Berry holds monthly internal meetings to track AI stylistic trends and update a running document ensuring the agency's copy doesn't drift toward LLM defaults.
In April, venture builder Ben Horwitz built Sinceerly, which he calls the anti-Grammarly. The app deliberately leaves typos in emails and signs off with sent from my iPhone. Horwitz A/B-tested standard emails against Sinceerly-generated ones by sending both to Fortune 500 CEOs; the only ones the executives opened were the lowercase, misspelled versions. Horwitz says the resulting stylistic homogeneity has made him really appreciate good writing more.
The theory behind the trend, at least in its most optimistic form, comes from John Warner, a veteran writing teacher who published More Than Words: How to Think About Writing in the Age of AI last April. Warner spent years trying to kill the five-paragraph student essay and what he calls pseudo-academic BS — the same kind of prose ChatGPT now produces on demand. His argument is that LLMs might reinstate the value of good writing by making the alternative unbearable to read.
“People now have a thirst for something different and for a lack of uniformity.”— Ben Horwitz, Venture builder, creator of Sinceerly
There is a real ceiling to all of this. Anti-AI stylistic markers — typos, unusual punctuation, first-person confessional openings — are themselves patterns, and patterns are what language models learn. McGarry says the point out loud: whatever cosmetic counterstyle is created in response to AI, AI could mimic, and the only true counterstyle is good writing. Once training corpora start absorbing the anti-AI style, the arms race resets one level up.
For AI companies, the counterstyle is a small but real signal about where the market for generative writing is heading. The volume side — LinkedIn posts, marketing copy, throwaway newsletters — is being commoditized by ChatGPT and Claude, and the human response is not to compete on volume but to compete on texture. Publishers like Playboy and Pitchfork are treating first-person authenticity as a commissioning filter, which effectively caps the addressable market for consumer AI writing tools inside serious editorial. The interesting play for model providers is not better em-dash suppression but better collaboration with human writers who want the model to disappear from the final draft — the opposite of what Grammarly and its imitators have spent a decade optimizing for.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
