The Enshittification of Writing

People who treat AI as a shortcut to replace reporting, sourcing, and thinking get churn that looks polished on the surface and brittle underneath. They’ve handed the machine the pen and decided supervision is optional. Writing with AI is fine when it’s used with judgment, verification, and accountability, the way I use it.

This one is ugly. The rot comes from people who ought to know better deciding that reading, checking and thinking were costs to cut, and the writers who still do that work are getting buried under a pile of confident filler. Readers pay for it one bad citation at a time, and most never learn which parts of what they read were invented.

You see the failure modes everywhere, starting with hallucinations, where models invent facts, quotes, or citations that don’t exist.

Plagiarism by remix, where outputs repurpose others’ structures or methods without clear attribution. Surface-level amplification, where content mills crank out thousands of barely varied posts that repeat the same ideas. These products crowd the web with material that reads like explanation but lacks experience.

Editors and researchers are documenting this pattern. Publishers that leaned too hard on AI have had to correct and retract pieces after readers caught errors and unattributed passages. That proves speed without oversight breaks credibility.

In academia, AI-generated manuscripts have been flagged for reusing others’ methods without citation, raising hard questions about credit and novelty. SEO practitioners warn that obvious AI patterns, including repetition, shallow coverage, awkward phrasing, and keyword stuffing, flood search results and push down original work. Detection and plagiarism tools help, but they’re imperfect and reactive, and they only catch damage that’s already done.

When Do Editing Tools Become the Problem?

Tools like Grammarly and ProWritingAid are useful when they catch typos, tighten sentences, or flag unclear phrasing. Misused, they accelerate enshittification. Push every sentence through automated “clarity,” “engagement,” and “readability” sliders and you sand off voice, flatten rhythm, and convert lived language into corporate mush. Over-reliance nudges writers toward safe synonyms, generic transitions, and passive constructions that score well in a dashboard but read like beige paste on the page. These apps are assistants. They’re great for mechanical cleanup and terrible as final editors of style, argument, or truth.

Autocrit takes the same problem and dials it up. Its reports tempt writers to chase “ideal” averages for pacing, repetition, and sentence mix until they hammer a living draft into template compliance. The result often reads like it was written to satisfy a spreadsheet: technically tidy, emotionally vacant, and indistinguishable from a thousand other drafts that hit the same metrics.

Used sparingly, these tools can flag issues worth a human decision. Used as the decider, they train writers out of voice and readers out of patience.

A grammar checker that catches typos is worth having. Writers and editors who let a score overrule their own ear infuriate me, because a dashboard can’t tell a flat sentence from a strange, brilliant one, and its makers built it to reward the flat one.

Watch: Cory Doctorow - Rescuing the Internet From “Enshittification” | The Daily Show, The Daily Show

Why This Matters

This matters because writing is how we share knowledge and make arguments. When content becomes churn, readers lose confidence. Brands that publish sloppy material sacrifice reputation. Research polluted by automated outputs steals attention and credit from the humans who did the labor of discovery.

I think the people hurt worst are the ones who never cut a corner: the researcher whose paper gets buried under generated commentaries, the freelancer underbid by a content mill, the reader who acts on a statistic somebody’s software made up. The people running the mills keep the ad revenue and walk away clean.

What Does the Enshittification of Writing Look Like?

The pattern plays out the same way across industries. A team under deadline pressure leans on AI to generate drafts, routes everything through grammar dashboards, and uses automated scoring as the final gate. Output looks successful at first. Traffic rises. Editorial calendars stay full. Then a reader tries to verify a citation and discovers the DOIs don’t resolve and the cited journal issues don’t exist. What began as a comment thread becomes a credibility problem.

The correction is usually quiet. Remove the bogus citations, edit the copy, skip the editor’s note. That choice costs more than expected. Conversion signals drop. Reliability questions spike in support queues and on social channels. Internally, the editor and writers spend days re-checking sources and patching links, erasing any short-term savings from automating verification. The reputational hit arrives as a slow leak: partners and prospects circulate the discovery in private messages, and the lead generation funnel stiffens until trust is rebuilt.

This pattern matches documented problems in academic publishing.

Predatory and low-quality journals have published AI-generated or misattributed articles containing fabricated references and DOI claims, forcing researchers and editors to police and retract content. The GIJIR investigation documented how an entire journal was populated with AI-generated articles falsely attributed to real researchers (Research Integrity and Peer Review). Neurosurgical Review retracted 129 papers after getting inundated by AI-generated commentaries and letters, most affiliated with a single university engaging in aggressive citation stacking (Retraction Watch).

AI-written preprints with fabricated references have been withdrawn only to resurface as preprints claiming to be under review elsewhere (Retraction Watch). The broader context is a rise in retractions and integrity failures tied to automation and low editorial vigilance (Nature).

Neurosurgical Review’s 129 retractions should have shamed the whole industry into slowing down. As far as I can see they didn’t, because the incentives that produced those papers are still paying out, and the researchers whose names got attached to fake work are the ones left explaining it.

How do you fix the enshittification of writing?

The remedy is practical and editorial. Keep humans central. Use AI for tasks it performs well, like brainstorming, outlining, or formatting, and never as a substitute for reporting, verification, or judgment. Require transparent attribution when AI plays a substantive role and document the verification steps taken. Train editors to spot AI fingerprints: repetitive phrasing, odd transitions, and unverifiable sources. Invest in fact-checking and source verification instead of chasing output quotas.

For academic work, peer review and novelty checks must adapt. Any AI-assisted manuscript should include provenance for ideas and explicit citations that can be verified. If methods overlap with prior work, reviewers should demand clear acknowledgment and treat uncredited idea reuse as a serious integrity issue.

Publishers and platforms should stop rewarding churn. If ad revenue and SEO tactics fund low-value content, the incentive to produce it stays in place. Cut the reward by prioritizing original reporting and verifiable data, require named authorship, and favor work that includes firsthand examples, interviews, or proprietary analysis.

Writers have to treat AI as a tool and do the authoring themselves. Add first-hand reporting, original examples and a distinct voice, and take responsibility for every claim you publish. If your name is on a piece, own the facts and the argument personally.

I tell anyone who asks me about AI the same thing. It’s a tool; treat it like one and you’ll do fine. Let it do your work for you and you’ll have problems, and every one of those problems will carry your name.

Three practical steps stop the damage when it starts.

First, require a verifiable primary source for every specific statistic or named study and log a timestamped verification step before publishing. Second, cut output targets so editors have time to do the work that automation cannot: calling sources, reading original papers, and confirming quotes. Third, publish transparent correction notices for substantive edits. Quiet fixes look like coverups and widen the trust gap.

Treating writing as a metric to optimize buys volume and spends trust, and I’ve lost patience with the publishers, platforms and content shops that keep making that trade with other people’s credibility. Verification is cheap next to a retraction. Anyone still skipping it to hit an output quota has decided readers matter less than a dashboard, and I’d expect those readers to stop coming back.

Schedule a free consultation to discuss your book project.

From the author

The Death of Thinking
Careless writing is one symptom of outsourced judgment, and I wrote The Death of Thinking about what that costs. It traces how reasonable shortcuts pile up into a pattern.

Frequently Asked Questions