Pickles Prague Festival: A Celebration of Pickled and Fermented Vegetables
Partner article
Sep 08, 2026
AI Companion Users Are More Likely to Call AI Romance Cheating, Survey Finds
Prague Morning
Follow us on Google Discover
Current users of AI girlfriend and boyfriend apps were nearly twice as likely as those who had never used one to say a partner’s AI romance could count as cheating, according to a 2026 survey of U.S. adults.
People who use AI companion apps might be expected to take a relaxed view of relationships with artificial intelligence. New survey data points the other way.
Among respondents currently using an AI girlfriend or boyfriend app, 80.6% said a partner’s romantic or sexual relationship with an AI could count as cheating. Among people who had never used one, 44.0% said the same. Former users sat between the two, at 60.2%.
The figures come from a 2026 survey on AI romance and modern relationships conducted by AI Girlfriend Coach, a site that tests AI companion apps, through SurveyMonkey Audience on August 26 and 27. It collected 2,150 responses from U.S. adults aged 18 and over; 1,709 were retained after a quality screen removed a cluster of patterned answers. All figures describe that screened sample rather than the U.S. population, and the study did not cover attitudes in Czechia or elsewhere in Europe.
Why the result stands out
The finding cuts against an obvious assumption: that the people spending the most time with AI companions would be the ones most likely to dismiss them as “only a chatbot.”
Instead, the people with no experience of the apps were the least likely to see AI romance as potential cheating, and the most likely to say the question does not apply because an AI is not a real person. A third of never-users, 33.0%, took that position, against 11.4% of current users. Current users’ single most common answer, chosen by 43.6%, was that a partner’s AI romance becomes cheating once they grow emotionally or sexually attached, an answer only 9.5% of never-users gave.
The survey shows an association, not a cause. It cannot say whether using these apps changes how seriously people take them, whether people who already take them seriously are more likely to try one, or whether something else explains the gap. What it does show is that firsthand experience was not linked to dismissing the question. If anything, the reverse.
Half the sample already saw a boundary
The finding arrives against a backdrop in which AI romance was already contested.
Across all 1,709 screened respondents, 50.5% said a partner’s romantic or sexual relationship with an AI could constitute cheating. That breaks down as 33.4% who said any such interaction counts and 17.1% who said it becomes cheating once emotional attachment develops. A further 21.2% would not call it cheating but would be uncomfortable with it, leaving 28.3% who said it is not cheating because an AI is not a real person.
Put another way, 71.7% of respondents expressed some level of discomfort with a partner having an AI romance, whether or not they were prepared to use the word cheating.
Relationship status mattered too. Engaged or married respondents were more likely than single respondents to call AI romance cheating, at 54.8% versus 42.2%, and singles were the group most likely to say an AI is not a real person.
Believing in AI feelings goes with stricter judgment
A second pattern in the survey points in the same direction.
Respondents were asked whether they could imagine developing genuine romantic feelings for an AI. Among the 500 who said yes or maybe, 73.0% considered a partner’s AI romance cheating. Among the 1,209 who said they could never have such feelings, 41.2% did.
The people most open to AI attachment were, in other words, the strictest judges of a partner’s AI attachment. That is the same direction as the user finding, and it suggests the divide in this survey is less about whether cheating requires another human than about whether respondents believe an emotionally meaningful relationship with software is possible at all. Those who believe it is tended to treat it as consequential.
Companionship, not sex, is the expected appeal
Part of what makes the boundary question difficult is that the survey found AI companionship is not, in most respondents’ minds, primarily about sex.
Asked what the main reason for using an AI companion would be, 56.6% chose having someone to talk to or feeling less lonely. Only 10.3% chose affection, romance or sexual interaction. The question was hypothetical and asked of everyone, so it reflects what respondents believed the appeal would be rather than the confirmed motives of users.
That matters for couples. A partner using an app for sexual roleplay is a relatively easy case to argue about. A partner who talks to an AI every night about work, worries and the relationship itself is a harder one, and the survey’s 17.1% who located cheating at the point of attachment suggest many people already sense the difference.
Not a replacement for people
None of this indicates that respondents, users included, saw AI as a substitute for human relationships.
Asked what they would choose if single and looking for companionship, 52.2% picked dating apps, 7.8% an AI girlfriend or boyfriend, and 13.2% both. And 64.6% said an AI could never replace a human relationship in their own lives.
A person can believe an AI cannot replace a human partner and still believe that a partner’s interactions with one can cross a line. The findings show that skepticism about AI replacing human relationships can coexist with widespread concern about AI romance crossing relationship boundaries.
What it means for couples
The practical upshot is that AI companionship is becoming one more boundary couples may need to negotiate, alongside private messages, close friendships and what stays on a phone after a relationship becomes serious. The survey does not offer a universal answer to where that boundary sits; its respondents disagreed among themselves. But it does indicate that the disagreement is not between people who take AI seriously and people who do not. Current users, the group actively using the technology, were the most likely to see a line to cross.
The survey does not explain why current AI companion users were almost twice as likely as never-users to say AI romance could count as cheating. It cannot determine whether the difference reflects who chooses to use these apps, the experience of using them, or other factors it did not measure. But the finding challenges one simple assumption: that people who spend time with AI companions necessarily regard those interactions as less meaningful. In this sample of American adults, current users were the group most willing to classify a partner’s AI romance as potential cheating.
Partner Content
Would you like your story featured here? Get in touch with our team at [email protected]
