Does polished, confident writing pass for real expertise, and can a structured check catch the difference?

This explores how far a polished surface, meaning fluent writing, professional formatting and a confident tone, can stand in for real expertise when readers judge quality or truth. The corpus's short answer: polish substitutes for expertise to a striking degree whenever judgment is left to gut feel. It substitutes much less when readers are given a structured way to check substance.

The evidence that polish wins is consistent. Evaluators rated AI-generated documents as both human-written and better than real human submissions Does polished writing actually signal better quality work?. Even readers with machine-learning expertise couldn't reliably tell LLM-written research abstracts from human ones, and they gave LLM-edited versions the highest clarity ratings Can readers tell LLM abstracts from human ones?. Models trained to imitate ChatGPT fooled human evaluators by copying its confident, fluent style while gaining nothing in factual accuracy Can imitating ChatGPT fool evaluators into thinking models improved?. In an 81-person study, readers with no information about where claims came from showed no measurable ability to tell truth from fabrication. They fell for fluent hallucinations as readily as for real facts Can readers tell truth from fabrication without evidence signals?. One explanation is that people have long relied on a reasonable shortcut: work that looks professional usually came from someone who thought hard. Generative AI breaks that link, and the break hurts newcomers most, because they have the least domain knowledge to look past the form Does polished AI output trick audiences into trusting it?.

Two findings push this further than most people would expect. First, the illusion also points inward. When users get polished AI output, they read its smoothness as evidence of their own competence, even though they didn't produce it Does processing ease mislead users about their own competence?. Second, machines fall for it too. LLMs used as judges reliably reward fake references and rich formatting, and these tricks work without any access to the model Can LLM judges be fooled by fake credentials and formatting?. So handing evaluation to AI doesn't remove the bias. It automates it.

The substitution has limits, though, and they show where it breaks. In 192 human-vs-LLM debates, LLMs won the crowd's preference votes but did noticeably worse under formal scoring of argument strength, where humans stayed competitive Do fluent arguments win debates through sound logic or rhetorical polish?. Persuasiveness and sound reasoning turn out to be separate skills, and polish only buys the first. Showing readers which claims are verified brought their truth-detection back, with a +4.15 point gap Can readers tell truth from fabrication without evidence signals?. Wording may also matter less than we assume. In debate data, readers' prior political and religious beliefs predicted who won better than any feature of the language did Does what readers believe matter more than what debaters say?.

The less obvious lesson concerns what polish is actually imitating. Several notes argue that expertise isn't mainly a stock of knowledge. It is a social performance: knowing when to speak, when to defer, and what a particular audience will accept Is expertise really just knowing more than others? Can AI replicate the communicative work experts do?. Polish copies the outward signs of that performance without the judgment that produced them. AI assistance even changes how readers see the writer, shifting all 29 measured traits toward seeming more confident, higher-quality and more privileged Does AI writing assistance change how readers perceive the writer?. In practice, polish now tells you little about who did the thinking or how well they did it.