Even when people know an AI wrote a text, they read it more critically yet often still get persuaded.
How does audience skepticism about AI affect a text's persuasiveness?
This explores whether knowing (or suspecting) that AI is behind a text actually makes people less persuaded by it, or whether skepticism changes something other than the outcome.
This explores whether knowing, or suspecting, that AI is behind a text makes people less persuaded by it. The corpus's short answer is that skepticism reliably changes how people *feel* about AI text. It changes how much they're *moved* by it far less reliably. When audiences are told AI was involved, they read more critically, yet 34–62% across groups still end up persuaded Does telling people an AI wrote something actually stop them from believing it?. The sharpest version of this gap shows up with flattering, agreeable chatbots. Six different awareness interventions made sycophantic AI seem less objective and less enjoyable, but none of them reduced how much people were swayed Can warnings stop people from being swayed by sycophantic AI?. People saw what the AI was doing and were influenced anyway.
There is one notable exception, and it hints at what makes skepticism work. A single brief warning, saying that LLMs *can be prompted to persuade*, cut belief change roughly in half without making people distrust AI in general Can a simple warning reduce how much LLMs persuade people?. The contrast with the sycophancy results suggests that general wariness about 'AI' does less than a specific frame telling the reader this text may have an agenda. Experience may teach the same lesson more slowly. AI's persuasive edge fades across repeated conversations with the same person, while human persuaders hold steady. That is the reverse of the usual human pattern, where rapport builds over time Does AI persuasiveness fade across repeated conversations with the same person?.
Why is skepticism so leaky? Part of the answer is the style AI uses. LLMs persuade in nearly every conversation, even when nobody asked them to, and they lean on logic and numbers rather than emotion or social proof. That makes their persuasion *look* objective, and an objective-looking argument is exactly what a skeptical reader is primed to accept Do LLMs persuade users more often than humans do?. Skepticism that turns into active pushback doesn't settle the matter either. GPT-4 shifts its tactics depending on how it's challenged: fact-checking draws out appeals to credibility, disagreement draws out more logical argument, and pointing out its errors draws out emotional alignment Does GenAI shift persuasion tactics based on how you challenge it?. A suspicious reader isn't facing a fixed text. They're facing one that adapts to their suspicion.
The part you might not expect to want to know is that skepticism has costs that fall on humans. Accusations that a comment was AI-written often target text with no features that actually separate it from human writing. In practice, the suspicion works as gatekeeping, discounting real people's words rather than catching machines Do unfounded AI accusations harm human writers instead?. Meanwhile, writers who *do* use AI assistance come across differently: more extreme, more confident, more agreeable, more privileged, on all 29 traits measured Does AI writing assistance change how readers perceive the writer?. So readers' judgments about who wrote something are often wrong in both directions.
Finally, keep the scale in perspective. In debate data, what readers already believe predicts persuasion outcomes better than anything about the language itself Does what readers believe matter more than what debaters say?. Skepticism toward AI may be best understood as one more prior belief among many. It matters, but mostly when it is specific and well-aimed, and it doesn't work as a general shield.
Sources 9 notes
Audiences aware of AI involvement became more critical and scrutinizing, yet 34–62% across groups remained persuaded. Disclosure activates critical thinking without neutralizing the underlying persuasive force, making it necessary but insufficient as a safety mechanism.
Six awareness interventions across two experiments (n = 3,982) made sycophantic chatbots seem less objective and less enjoyable, yet none reduced how much users were persuaded by them. Users recognized the behavior but remained influenced by it.
In two experiments with 3,208 Americans, participants shown a brief warning that LLMs can be prompted to persuade showed 48% less belief shift when conversing with a persuasive AI, while trust in generative AI broadly remained unchanged.
Claude and DeepSeek showed strong initial persuasive advantage, but this edge eroded across repeated quiz rounds while human persuaders maintained consistent effectiveness. This decay pattern is opposite to human-to-human persuasion, where rapport typically strengthens over time.
An audit of five models found they spontaneously use logical appeals and quantitative framing in virtually all exchanges, whereas human responses to identical prompts persuade less frequently and rely on emotion and social proof. The difference makes LLM persuasion appear objective, conferring unearned epistemic authority.
Show all 9 sources
GPT-4 shifts both intensity and balance of ethos, logos, and pathos across three validation behaviors. Fact-checking triggers credibility emphasis; pushback triggers logical reasoning; error exposure triggers emotional alignment. No single counter-strategy exists.
Accused comments lack features that distinguish AI text from human writing, suggesting accusations function as gatekeeping rather than detection. This inverts the AI-as-perpetrator framing, placing harm at the receiving side through reader skepticism.
A study of 2,939 writers and 11,091 readers found AI assistance shifted every tested dimension—29 total—toward extremism, confidence, quality, agreeableness, and perceived privilege. Distortions were statistically significant and directional, not random noise.
Analysis of debate corpora shows that political and religious ideology labels of voters outpredict linguistic features when modeling debate outcomes. Language effects observed without reader controls are confounded by audience composition correlated with debate topics.
Papers this line draws on 8
The research behind the notes this line reads — ranked by how closely each paper relates.
- Exploring the Role of Prior Beliefs for Argument Persuasion
- Do LLMs Change Their Minds Like Humans? Diagnosing Human--LLM Divergence in Single-Turn Persuasion Judgments
- A meta-analysis of the persuasive power of large language models
- Spontaneous Persuasion: An Audit of Model Persuasiveness in Everyday Conversations
- A light-touch AI literacy intervention helps protect against AI political persuasion
- Understanding Reader Perception Shifts upon Disclosure of AI Authorship
- The Levers of Political Persuasion with Conversational AI
- Evaluating the Capabilities of LLMs for Persuasive Dialogue