Do people trust emails they don't know were AI-assisted, and will that trust turn into suspicion as awareness of AI grows?
Will recipient skepticism of unlabeled AI messages grow as AI awareness increases over time?
This explores whether people's tendency to trust AI-written messages they don't know are AI-written will wear off as AI tools become more widely known, and what that growing suspicion might actually look like.
This explores whether readers' easy trust in unlabeled AI messages will turn into suspicion as people become more aware of AI-generated text. The honest answer is that the corpus can't yet show this happening. The key experiment (N=647) found that recipients rated unlabeled AI-assisted emails just as favorably as human-written ones. Strong skepticism appeared only once the AI's involvement was disclosed Do readers trust unlabeled AI-written messages as much as human ones?. The authors expect that rising awareness may wear down this default trust. But they also say plainly that a one-time snapshot can't measure whether that erosion happens Does trust in unlabeled AI messages decline as awareness grows?. So the prediction is plausible, but nobody has tracked it over time yet.
The nearby research says more about what kind of skepticism awareness produces than about whether it grows. Telling audiences that AI was involved does make them more critical, yet 34–62% stayed persuaded anyway Does telling people an AI wrote something actually stop them from believing it?. A brief, general warning that LLMs can be prompted to persuade cut belief change roughly in half, and it didn't lower people's overall trust in AI Can a simple warning reduce how much LLMs persuade people?. Taken together, general awareness seems to work more like a speed bump than a wall. It makes people more careful without making them reject AI messages outright.
The less obvious risk is that growing suspicion may land on the wrong targets. One study of accusations of AI use found that the accused comments had none of the features that actually set AI text apart from human writing. The accusations worked more as gatekeeping than as detection, so the harm fell on human writers who were wrongly doubted Do unfounded AI accusations harm human writers instead?. The social stakes are real: about half of people who received sloppy AI-generated work ('workslop') rated the sender as less capable, and 42% saw them as less trustworthy Does receiving AI-written work change how we judge the sender?. As awareness rises, readers may get more suspicious without getting any better at spotting AI. Suspicion that can't tell AI from human writing ends up taxing everyone.
Time may also push in the opposite direction. When people knew they were dealing with an AI partner, they at first avoided it. That bias reversed after repeated interactions in which they could see the results. The key was the feedback, not the label Does revealing AI identity help or hurt user trust?. That suggests a different path over the long run: a wave of suspicion as awareness spreads, followed by trust that settles according to whether AI-assisted messages actually prove reliable. So the better question may not be whether skepticism will grow. It may be whether readers will get the outcome feedback they need to turn vague suspicion into well-judged trust.
Sources 7 notes
In a preregistered experiment (N=647), recipients rated unlabeled AI-assisted emails indistinguishably from human-written ones. Only explicit AI disclosure triggered strong skepticism. Recipients appear to default to trust rather than suspicion when origin is unrevealed.
In a single study of 647 participants, readers rated unlabeled AI-assisted messages as favorably as human-written ones. The authors predict awareness may shift this baseline but acknowledge their snapshot design cannot measure whether that erosion actually occurs.
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.
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.
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.
Show all 7 sources
About half of survey respondents who received workslop rated the sender as less creative, capable, and reliable. Forty-two percent viewed them as less trustworthy, and nearly one-third said they'd be less willing to work with them again.
Users initially avoid AI partners when identity is revealed, but this preference reverses after repeated interactions with visible results. The learning mechanism—observing consistent outcomes—is essential; disclosure without feedback produces no calibration.
Papers this line draws on 8
The research behind the notes this line reads — ranked by how closely each paper relates.
- Understanding Reader Perception Shifts upon Disclosure of AI Authorship
- Penalizing Transparency? How AI Disclosure and Author Demographics Shape Human and AI Judgments About Writing
- Toward Meaningful Transparency for AI Chatbots: Disclosing Persuasive Intent Reduces Persuasion
- Blissful (A)Ignorance: People form overly positive impressions of others based on their written messages, despite wide-scale adoption of Generative AI
- A light-touch AI literacy intervention helps protect against AI political persuasion
- What Influences Readers' and Writers' Perceived Necessity of AI Disclosure?
- The human-authorship halo: attribution bias in literary style evaluation by humans and AI
- Humans learn to prefer trustworthy AI over human partners