When AI polishes your writing, do readers start picturing a different person behind the words, with more schooling and money?

This explores whether AI writing help changes who readers think the writer is, and not only how polished the text sounds. The corpus answers yes, and the effect is large. In a study of nearly 3,000 writers and 11,000 readers, AI assistance shifted all 29 measured traits of the perceived author. Readers judged the writers as more confident, more agreeable, more extreme in their views, and more privileged than they judged the same writers' unassisted text Does AI writing assistance change how readers perceive the writer?. On demographics, readers were 5.3× more likely to judge AI-assisted writers as highly educated, 4.4× more likely to judge them as higher-income, and 4.1× more likely to judge them as native English speakers. The researchers call this 'identity laundering' Does AI writing make authors seem more privileged than they are?.

The less obvious finding is that AI doesn't just make each writer look different. It makes them look the same. Variation in perceived author traits narrowed on 22 of the 29 dimensions, so writers converged on one confident, articulate persona Does AI writing make all writers sound the same?. That persona has a cultural direction. In a separate experiment, GPT-4o autocomplete pulled Indian writers toward Western phrasing and cultural references, and it also gave American users bigger productivity gains. So the writers whose voices get overwritten most are also the ones who benefit least Do AI writing assistants push non-Western writers toward Western styles?. Computational authorship tools find the same erasure. Heavy AI rewriting cut author-identification accuracy by 66.5 points on personal blogs but only 10 points on news writing. Personal, voice-driven writing loses the most How much does AI rewriting erase distinctive author voice?.

The distortion reaches readers because writers rarely edit it out. They changed AI-generated paragraphs only 23% of the time, and edited versions stayed about 96% similar to the original Do writers actually edit AI-generated text before publishing?. Fixing this is hard. When researchers trained models to reduce the persona distortion, writers liked the output less. The clarity and confidence people want from AI seem to come from the same tendencies that produce the distortion Can AI writing assistance remove distortion without losing appeal?. One related point: people feel more ownership of AI text when they have more control over what it says, while personalizing the model to them doesn't change that feeling Does user control over AI text shape feelings of ownership?. Making the model 'sound like you' may matter less than letting you steer it.

Reader perception also runs the other way. When readers know or suspect AI was involved, their judgments shift again. Identical passages were rated 13.7 points lower when labeled as AI-written, and AI judges showed an even stronger bias of 34.3 points Do authorship labels bias how we judge literary quality?. Disclosing AI use cut perceived trust and caring most in personal, interpersonal writing How does revealing AI authorship change reader trust?. Suspicion can also fall on people who didn't use AI at all. Accusations of AI use often target comments that have no actual AI markers, which makes them a form of gatekeeping that wrongly discredits human writers Do unfounded AI accusations harm human writers instead?. Taken together, writers face a squeeze. With AI help they come across as a generic privileged persona. Writing in their own voice, they may be suspected of using AI anyway, and the writers most affected on both sides are often the ones whose voices were already least typical of the model's training data.