AI scams erode trust in online identity, Malwarebytes warns

Malwarebytes has published research on how artificial intelligence is affecting trust, scams and online identity. The survey found that one in three daily AI users think it is acceptable to create explicit images of people they know.

The report drew on responses from 1,500 adults in the US, UK, Austria, Germany and Switzerland, and pointed to growing uncertainty over whether online material is real and communications are genuine.

One of the clearest findings was a decline in confidence in digital evidence. Some 88% of respondents said it is becoming harder to tell whether online content is genuinely human or real, while 84% said convincing video evidence no longer feels like proof.

Scams were another major concern. Some 85% of respondents said they struggle to distinguish scams from legitimate communications, up from 66% the previous year.

Half of those surveyed said they had experienced some form of AI fraud or scam. Exposure was highest among Gen Z respondents at 67%, compared with 51% of Millennials, 46% of Gen X and 30% of Boomers and older people.

The data also suggested identity-related abuse is becoming more common. One in 10 respondents said explicit AI images had been made of them without consent, while 19% said they had experienced some form of AI-driven identity harm. That figure rose to 30% among Gen Z.

Trust erosion

The research described a broad weakening of confidence in basic online signals such as voice, image and video. It found that AI-generated deepfakes, voice cloning and impersonation are contributing to what Malwarebytes characterised as a breakdown in certainty over what people can trust.

Regional differences also emerged. The US recorded higher exposure to AI fraud and scams at 56%, compared with 48% in the UK and 47% across the DACH region.

At the same time, concern was not always matched by defensive action. While 81% of respondents said they fear someone stealing their family’s likeness, only 13% said they had created a family codeword as a safeguard.

Similarly, 67% said they worry about voice cloning, but only 19% said they had turned off voicemail recordings to reduce that risk. The findings also showed that 74% are concerned about experiencing a deepfake or other AI-generated scam.

The DACH region lagged the US and UK across most protective behaviours measured in the study. The report suggested this may reflect stronger institutional trust in those markets.

Changing norms

Beyond fraud, the survey pointed to a shift in attitudes about what people consider acceptable AI use. It found that 18% of respondents believe it is acceptable to use AI to generate explicit images of someone they do not know.

Among daily AI users, the picture was more striking. One in three said it is acceptable to generate explicit images of someone without their consent.

Another 32% of respondents said it is acceptable to use AI to imitate their voice or appearance, provided it is for personal use. The findings suggest concern about misuse can coexist with tolerance for practices that could enable abuse.

Mark Beare, Head of Consumer at Malwarebytes, commented on the findings.

“AI’s deepest impact isn’t on our devices; it’s on us. When people can no longer trust what they see, hear, or who they’re talking to, the damage reaches far beyond any single scam and into the building blocks of our society,” Beare said.

He also linked the issue to the wider role of cyber protection.

“Cybersecurity has always adapted, and it will again, but only if we recognize that what we’re protecting now is something far more important than data. It’s people’s ability to believe one another,” he said.

The report was based on a survey prepared by an independent research consultant and distributed through Forsta. Respondents were aged 18 and older, with the sample split equally by gender and weighted across age groups, regions and race groups.

Malwarebytes also used the publication of the findings to highlight Scam Guard, a scam-detection feature built into its desktop and mobile products. The tool provides real-time feedback on suspected scams, threats and malware, alongside digital safety recommendations.

It is also intended to reduce the stigma that can surround scam victims by offering guidance before users act on suspicious messages. The wider findings, however, indicate that the challenge may extend beyond technical detection to a deeper loss of confidence in whether online interactions can be trusted at all.