Why do people see AI-assisted work as less legitimate, and is the doubt really about quality or about who did it?

This explores why AI-assisted work gets treated as less legitimate: who applies the penalty, what it costs, and whether the suspicion points at something real. The corpus clearly shows that the penalty exists. It says less about the psychology behind it. What it does offer is a surprising angle: the doubt may have more to do with authorship and effort than with quality.

Start with the social cost, because people see it coming. In experiments with more than 4,000 participants, people who used AI expected colleagues to rate them as less competent and less diligent, so they were less willing to say they had used it Do people fear judgment when they use AI at work?. The fear seems justified. When people received low-effort AI-generated work ('workslop'), about half rated the sender as less creative, capable, and reliable, and 42% trusted them less Does receiving AI-written work change how we judge the sender?. Note the word 'diligent': the judgment seems to be about effort, not only about how good the result is.

This is where things get odd. People can't actually tell AI work from human work. A review of 30 studies found that detection across text, images, and voice is roughly a coin flip Can people reliably spot content made by AI?. That doesn't mean there are no differences. Across thousands of stories, AI fiction reliably over-explains its themes and avoids ambiguity Do AI stories explain their themes more than human stories do?. Those patterns show up in aggregate, though, and they're too weak to judge a single piece by. So the penalty often hits based on suspicion rather than evidence. One study of online comments found that the ones accused of being AI had no features that set them apart from human writing. The accusations worked as gatekeeping, and the people harmed were human writers who weren't believed Do unfounded AI accusations harm human writers instead?.

There's an uncomfortable twist. The doubters may be picking up on something real, just not the thing they think. Several notes describe the 'LLM Fallacy': users come to believe that AI-assisted output shows skills they don't actually have Do AI-assisted outputs fool users about their own skills?. This is a mistake about one's own abilities, separate from hallucination or over-trusting the AI How does AI-assisted work reshape how people see their own abilities?. Users claim authorship socially without the cognitive ownership that would normally come with it. The cause is not dishonesty. The intermediate steps are hidden, and people build a story about their role afterward Do users truly own the AI-generated content they produce?. Four mechanisms feed this and amplify each other: unclear credit, the polish of fluent output, handing off the thinking, and an opaque process How do AI tools trick users into overestimating their own skills?. Over time, leaning on AI can wear down the skills that oversight depends on Does AI augmentation protect workers from skill erosion?.

The takeaway you may not have expected: 'Is this AI?' is mostly a question nobody can answer, and asking it often wrongs human authors. 'Does the person behind this understand and stand behind it?' is a question about legitimacy that the research suggests is reasonable to ask. Even AI users themselves often get it wrong. The corpus doesn't directly test why observers apply the penalty, for example whether they use visible effort as a stand-in for merit or worry about fairness. That part is still open.