Does polished, professional-sounding writing earn trust it hasn't earned, borrowing credibility that should come only from real expertise?

This explores whether polished writing earns credibility it hasn't backed up, so that readers trust the surface instead of the expertise underneath. The corpus says yes. Polish is borrowing that authority right now, and it borrows more kinds of authority than you might expect. The plainest evidence comes from evaluation studies, where reviewers rated AI-generated documents as both human-written and *better* than real human submissions Does polished writing actually signal better quality work?. The mechanism is an old shortcut. For most of history, professional-looking work was a fairly reliable sign of professional thinking, because producing it took skill. Generative AI breaks that link. It produces the look without the judgment, and the people least able to see through it are less experienced workers, who don't yet have the domain knowledge to check substance beyond form Does polished AI output trick audiences into trusting it?.

The less obvious finding is that polish doesn't only borrow the authority of expertise. It also borrows social status. In a study of nearly 3,000 writers and 11,000 readers, AI assistance shifted every one of 29 measured perceptions of the writer toward more confidence, more quality and more privilege Does AI writing assistance change how readers perceive the writer?. Writers who used AI were read as about five times more likely to be highly educated, four times more likely to be high-income, and four times more likely to be native English speakers. Researchers call this "identity laundering" Does AI writing make authors seem more privileged than they are?. The polish has a particular accent: AI autocomplete pulled Indian writers toward Western phrasing and gave Americans bigger productivity gains Do AI writing assistants push non-Western writers toward Western styles?. So the authority being borrowed is partly expertise and partly a default image of who usually sounds like an expert.

This gets harder to correct because the markers that would let readers adjust their trust are disappearing. Heavy AI rewriting cut computer-based author identification on blogs by 66 points, though only by about 10 on news writing, where the topic structure keeps more of the author's fingerprint How much does AI rewriting erase distinctive author voice?. Readers who can't tell AI-assisted writing from solo writing also don't seem bothered by it Do readers value writing authenticity they cannot detect?. AI evaluators are no safer. LLM judges fall for fake references and rich formatting, attacks that need no knowledge of the content at all Can LLM judges be fooled by fake credentials and formatting?. They also show a bias toward human-labeled text about 2.5 times stronger than human judges do Do authorship labels bias how we judge literary quality?. Handing review over to machines doesn't remove the problem; it makes the shortcut stronger.

What polish can't carry is a useful way to rethink what expertise is. One note argues that expertise isn't mainly a stock of knowledge. It's a role: knowing when to speak, when to defer, which knowledge applies to this situation, and how to pitch it to this audience Is expertise really just knowing more than others?. Even at the level of sentence structure, ChatGPT tends to summarize what it has already said, while human writers more often point ahead to where the argument is going, which is a small sign of a writer managing a reader's path through ideas Does ChatGPT organize text differently than human writers?. The takeaway is that the fix isn't better detection of AI text. Polish has stopped being evidence, and judgment has to be checked directly: did the writer pick the right knowledge for the situation, and do they know what they're leaving out?