New ReverseLookup data suggests that confidence in spotting synthetic media often exceeds people’s willingness to acknowledge their own mistakes. The findings show how AI detection is becoming tied to reputation, social status and the wider question of whether digital evidence can still be trusted.
ReverseLookup surveyed 7,850 adults across the United States, Europe and Latin America about their confidence in identifying AI-generated images and videos, how they respond after making mistakes and which groups they believe are most vulnerable to synthetic content.
The results reveal a clear imbalance in how people assess their own abilities. 72% of respondents said they feel confident in their ability to recognize AI-generated content. When comparing themselves with the average internet user, 61% said they believe they are better at spotting synthetic media, while only 9% considered themselves worse than average.
That distribution points to more than widespread confidence. It suggests that many users overestimate their relative ability to identify synthetic content. Although the survey did not test respondents through a controlled image-recognition exercise, the gap between those who place themselves above average and those who place themselves below it shows how readily people recognize digital vulnerability in others while discounting it in themselves.
The findings also indicate that AI detection has become connected to personal reputation. Identifying manipulated content can function as evidence of digital competence, while failing to recognize it can feel like a public lapse in judgment. On social platforms, labeling an image or video as artificial allows users to present themselves as cautious and technically aware. Believing or sharing synthetic content carries the opposite implication.
That pressure becomes visible when people are asked how they would respond to an error. After realizing they had shared AI-generated content believing it was real, 43% said they would quietly delete the post. Only 31% said they would openly acknowledge the mistake.
The preference for deletion matters because it conceals how often confident judgments fail. When incorrect assessments disappear without explanation, other users see fewer examples of people being deceived. That can reinforce the impression that identifying AI-generated content is easier and more reliable than it actually is.
Reputation appears to weigh almost as heavily as accuracy. 47% of respondents said they would feel more uncomfortable about other people knowing they had believed AI-generated content than about the deception itself. For these users, the concern is not simply that a false image or video appeared credible. It is that believing it may change how others evaluate their judgment.
Respondents also tended to locate susceptibility outside their own demographic group. Only 18% identified their own age group as the one most likely to believe AI-generated content. Younger participants more often pointed to older users, while adults aged 35 to 44 were more likely to identify younger internet users as the most vulnerable.
The pattern does not establish that one generation is consistently better at detecting synthetic media. Instead, it shows how users interpret digital risk through assumptions about other groups. Older people may be associated with limited familiarity with emerging tools, while younger users may be seen as reacting too quickly to content distributed through fast-moving platforms. In both cases, vulnerability is treated primarily as someone else’s problem.
Overconfidence can also produce the opposite error: dismissing authentic content as artificial. 49% of respondents said people online are now too quick to label unusual or highly polished images and videos as AI-generated.
The survey also examined which visual details most often trigger suspicion. 58% said unusually smooth skin or faces were among the strongest signs that an image had been generated by AI, while 46% pointed to distorted hands or fingers. Another 41% associated highly symmetrical compositions with synthetic content, and 37% said images that appeared excessively polished or cinematic were more likely to be artificial.
These cues are not reliable on their own. Professional editing, beauty filters, video compression, unusual lighting and high-end photography can create many of the same effects. When users treat one visual irregularity as proof of AI generation, skepticism can become another form of misidentification.
This creates a double failure of digital judgment. Users may accept synthetic content because they overrate their ability to recognize manipulation, but they may also reject real material because identifying something as AI has become a visible signal of awareness. Confidence is not necessarily producing better verification. In some cases, it is encouraging faster and more public conclusions.
The consequences extend beyond individual embarrassment. When authentic photographs or videos are dismissed as AI-generated, genuine evidence can lose credibility before it is properly examined. Images documenting breaking news, public incidents or personal experiences may be rejected not because they have been disproved, but because they look unusual, emotionally charged or inconsistent with what viewers expect.
The wider assumption that convincing images can be fabricated also makes it easier to undermine real material without demonstrating that it has been manipulated. Synthetic media does not only add false content to the information environment. It also provides a ready-made reason to distrust content that may be genuine.
The ReverseLookup findings suggest that AI literacy is becoming partly performative. People want to believe they are difficult to deceive, prefer to conceal evidence that contradicts that belief and frequently assume other groups are more susceptible than their own. As synthetic media becomes more common, the central risk is not only that users will believe false content. It is that confidence in detecting it may weaken trust in real evidence as well.
About ReverseLookup
ReverseLookup is a multi-input verification platform for phone numbers, emails, and images. Built for everyday use, ReverseLookup.com enables users to assess unfamiliar contacts, investigate questionable profiles, and identify potential fraud across key digital channels. It combines reverse search methods with open-source intelligence (OSINT) to offer a direct, accessible way to review digital identities and make informed decisions online.
Media contact
Ashleigh Thomas
PR Manager
ReverseLookup
pr@reverselookup.com
The results reveal a clear imbalance in how people assess their own abilities. 72% of respondents said they feel confident in their ability to recognize AI-generated content. When comparing themselves with the average internet user, 61% said they believe they are better at spotting synthetic media, while only 9% considered themselves worse than average.
That distribution points to more than widespread confidence. It suggests that many users overestimate their relative ability to identify synthetic content. Although the survey did not test respondents through a controlled image-recognition exercise, the gap between those who place themselves above average and those who place themselves below it shows how readily people recognize digital vulnerability in others while discounting it in themselves.
The findings also indicate that AI detection has become connected to personal reputation. Identifying manipulated content can function as evidence of digital competence, while failing to recognize it can feel like a public lapse in judgment. On social platforms, labeling an image or video as artificial allows users to present themselves as cautious and technically aware. Believing or sharing synthetic content carries the opposite implication.
That pressure becomes visible when people are asked how they would respond to an error. After realizing they had shared AI-generated content believing it was real, 43% said they would quietly delete the post. Only 31% said they would openly acknowledge the mistake.
The preference for deletion matters because it conceals how often confident judgments fail. When incorrect assessments disappear without explanation, other users see fewer examples of people being deceived. That can reinforce the impression that identifying AI-generated content is easier and more reliable than it actually is.
Reputation appears to weigh almost as heavily as accuracy. 47% of respondents said they would feel more uncomfortable about other people knowing they had believed AI-generated content than about the deception itself. For these users, the concern is not simply that a false image or video appeared credible. It is that believing it may change how others evaluate their judgment.
Respondents also tended to locate susceptibility outside their own demographic group. Only 18% identified their own age group as the one most likely to believe AI-generated content. Younger participants more often pointed to older users, while adults aged 35 to 44 were more likely to identify younger internet users as the most vulnerable.
The pattern does not establish that one generation is consistently better at detecting synthetic media. Instead, it shows how users interpret digital risk through assumptions about other groups. Older people may be associated with limited familiarity with emerging tools, while younger users may be seen as reacting too quickly to content distributed through fast-moving platforms. In both cases, vulnerability is treated primarily as someone else’s problem.
Overconfidence can also produce the opposite error: dismissing authentic content as artificial. 49% of respondents said people online are now too quick to label unusual or highly polished images and videos as AI-generated.
The survey also examined which visual details most often trigger suspicion. 58% said unusually smooth skin or faces were among the strongest signs that an image had been generated by AI, while 46% pointed to distorted hands or fingers. Another 41% associated highly symmetrical compositions with synthetic content, and 37% said images that appeared excessively polished or cinematic were more likely to be artificial.
These cues are not reliable on their own. Professional editing, beauty filters, video compression, unusual lighting and high-end photography can create many of the same effects. When users treat one visual irregularity as proof of AI generation, skepticism can become another form of misidentification.
This creates a double failure of digital judgment. Users may accept synthetic content because they overrate their ability to recognize manipulation, but they may also reject real material because identifying something as AI has become a visible signal of awareness. Confidence is not necessarily producing better verification. In some cases, it is encouraging faster and more public conclusions.
The consequences extend beyond individual embarrassment. When authentic photographs or videos are dismissed as AI-generated, genuine evidence can lose credibility before it is properly examined. Images documenting breaking news, public incidents or personal experiences may be rejected not because they have been disproved, but because they look unusual, emotionally charged or inconsistent with what viewers expect.
The wider assumption that convincing images can be fabricated also makes it easier to undermine real material without demonstrating that it has been manipulated. Synthetic media does not only add false content to the information environment. It also provides a ready-made reason to distrust content that may be genuine.
The ReverseLookup findings suggest that AI literacy is becoming partly performative. People want to believe they are difficult to deceive, prefer to conceal evidence that contradicts that belief and frequently assume other groups are more susceptible than their own. As synthetic media becomes more common, the central risk is not only that users will believe false content. It is that confidence in detecting it may weaken trust in real evidence as well.
About ReverseLookup
ReverseLookup is a multi-input verification platform for phone numbers, emails, and images. Built for everyday use, ReverseLookup.com enables users to assess unfamiliar contacts, investigate questionable profiles, and identify potential fraud across key digital channels. It combines reverse search methods with open-source intelligence (OSINT) to offer a direct, accessible way to review digital identities and make informed decisions online.
Media contact
Ashleigh Thomas
PR Manager
ReverseLookup
pr@reverselookup.com