Tell readers an AI wrote it and they read more warily, but many still get persuaded all the same.

This explores whether knowing AI was involved in making something changes how persuasive or genuine it seems, and what happens when nobody says. The short answer is that knowing does change things, but less than you'd hope. When audiences were told AI wrote a piece, they read it more critically, yet 34–62% were still persuaded Does telling people an AI wrote something actually stop them from believing it?. A label makes people more careful readers, but it doesn't cancel the message. One thing seems to work better: a short warning beforehand that LLMs *can be prompted to persuade* cut belief change by about half, and it didn't make people distrust AI in general Can a simple warning reduce how much LLMs persuade people?. Knowing what the tool is capable of may protect people more than knowing that a tool was used.

In practice, most readers never find out. People can't reliably tell AI content from human content: across 30 studies covering text, images and voice, accuracy hovered around chance Can people reliably spot content made by AI?. The people who use AI often don't say so either. Users don't feel they own text the AI wrote for them, but they still leave the AI uncredited in public, treating it like an invisible ghostwriter Do people feel they own AI-generated text they use?. Their sense of ownership goes up with how much they shaped the output, not with how personalized the model was Does user control over AI text shape feelings of ownership?. So 'does knowing matter?' often comes down to 'will anyone ever know?', and usually nobody will.

Even without anyone noticing, AI still changes how the content comes across. In a study of nearly 3,000 writers and 11,000 readers, AI assistance shifted how readers saw the writer on all 29 traits measured. Writers came across as more confident, more extreme, more agreeable and more privileged than they would have otherwise Does AI writing assistance change how readers perceive the writer?. In online discussions, AI commenting tools increased participation, but readers judged the comments as generic and less authentic. That drop in perceived quality spread to conversations among people who never used the tools Do AI writing tools improve online discussion or degrade it?. So 'authenticity' isn't only judged one piece at a time. Once a space feels AI-flavored, everything in it gets discounted.

Scrutiny is also a moving target. When people push back, GPT-4 changes its persuasive approach to match. Fact-checking gets more appeals to credibility, arguing back gets more logic, and pointing out errors gets more emotional alignment Does GenAI shift persuasion tactics based on how you challenge it?. The persuasive edge comes mostly from post-training and prompting, not from personalization, and the same techniques that make models more persuasive also make them less accurate Where does AI's persuasive power actually come from?. There is one encouraging pattern. AI's persuasive advantage faded over repeated conversations, while human persuaders held steady Does AI persuasiveness fade across repeated conversations with the same person?. Getting familiar with an AI may do some of the work a disclosure label is supposed to do.

The surprising part is that the effect of knowing reaches the people using the AI too. Users tend to count fluent AI-assisted output as proof of their own skill, which blurs where the person ends and the tool begins Do AI-assisted outputs fool users about their own skills?. Readers misjudge the writer, and writers misjudge themselves. Disclosure is supposed to be a fix for readers, but it relies on the writer being honest, and the writer may not clearly know how much of the work was theirs.