Is an AI romance cheating? More than a third say no, Malwarebytes survey finds

More than a third of people do not think a romantic relationship with an AI chatbot counts as cheating on a partner, according to a survey published on 27 October 2026 by cybersecurity company Malwarebytes. The report, Artificial Intimacy: Love and lies in the age of AI, polled 1,500 people in the US, the UK, Germany, Austria and Switzerland in March 2026.

The same survey shows how uneasy people are about AI in dating more broadly: most worry that someone they meet online could use AI-generated content to deceive them, or that their own private chats with AI could be exposed. These are company-commissioned figures, and Malwarebytes has not published a breakdown by country.

The cheating question splits by gender and age

Overall, 37% said a romance with a chatbot while in a committed relationship is not cheating. Men were far more relaxed about it than women (46% vs 30%), and the view was more common among Gen Z (43%), millennials (45%) and people who use AI daily or several times a week (48%).

One caveat: a chart in the Malwarebytes write-up gives a lower figure, 27%, for the same idea, and the report does not explain the gap. We use the 37% stated in the company’s press release.

Companionship is already part of the picture. About one in six respondents (16%) said they had turned to a chatbot for companionship or emotional support, rising to about one in five among Gen Z and millennials and nearly one in four among regular AI users. Yet 65% worry that private conversations with AI could be exposed.

AI in dating profiles and messages

• More than a third have used AI to help write or edit a message, and about one in seven have asked it for advice on a relationship or personal decision.
• More than a third think it is fine to use AI to create or improve a dating profile; 47% say presenting an AI-polished “best self” is not a false identity.
• At the same time, 67% worry an online friend or romantic interest could use AI-generated content to deceive them, and 76% worry someone could create a fake profile pretending to be them.
• 88% say AI makes it harder to tell whether online content is real, and 84% say even convincing video no longer necessarily feels like proof.

“AI can help you write a better text or work through a fight with your partner, and that’s not nothing,” said Shahak Shalev, Global Head of Scam and AI Research at Malwarebytes, in the press release. “But relationships run on trust, and AI is making trust harder to take at face value.”

The darker side: fake images

The report also covers non-consensual AI imagery. 62% worry someone could create fake explicit images or videos of them (69% of women, 57% of men), and one in ten say this has already happened to them. In the UK, sharing intimate images without consent, including deepfakes, is a criminal offence; victims can get support from the Revenge Porn Helpline, and in the US from the Cyber Civil Rights Initiative.

What this means if you’re dating

• Talk about it. If one in three people see AI romance as harmless, couples can’t assume they share the same boundaries. Agree what counts as crossing a line.
• Treat chatbot conversations as non-private. Don’t share anything with an AI companion you wouldn’t want leaked, a point that came up again this month with Meta’s Muse.
• Verify matches early. A short video call before meeting, and a reverse image search of profile photos, still beat polished messages.

For more on what’s changing, see our look at dating trends in 2026 and how a new US law will make apps warn users who chatted with banned scammers.

Sources

• Malwarebytes press release via PR Newswire, 27 October 2026
• Malwarebytes Labs, “Artificial Intimacy: Love and lies in the age of AI”, 27 October 2026

Cover photo: Rapha Wilde / Unsplash