Get the app
How it worksThe castCompareGuidesPricingFAQ Languages About

Guide

Is an AI girlfriend cheating? How couples and researchers draw the line

No survey has settled it, and most users say it is not. The same users mostly hide it: 69% of partnered young adults said it mattered that their partner not learn the full extent.

By LORAX Team Published Updated

No survey has settled it, and most people who do it say it is not cheating. The finding that keeps recurring is about concealment rather than definition: among partnered young adults who regularly use an AI romantic companion, most have not told their partner the whole of it.

That gap is the useful part. A question nobody can answer for you sits next to a behavior that has been measured precisely. If you are here because a partner mentioned one and you are not sure what the thing even is, what an AI girlfriend actually is covers that first.

What the research actually measured

In May 2026 the Wheatley Institute at Brigham Young University and the Institute for Family Studies published a survey of 2,431 US young adults who were dating, engaged or married. One in seven, 15%, said they regularly interact with an AI chatbot that simulates a committed romantic partner. Another 20% to 30% said they had experimented with one at some point (Wheatley Institute and Institute for Family Studies, 2026).

Then the survey asked who knew.

Thirty percent of regular users said their real-life partner had no knowledge of it. Eleven percent said their partner was somewhat aware. Fourteen percent said mostly aware but not fully. Of those whose partners did know something, respondents estimated their partner was aware of roughly 37% of their use, and only 28% said their partner knew about more than half of it. Sixty-nine percent said it was somewhat or extremely important to them that their partner not learn the full extent (Wheatley Institute and Institute for Family Studies, 2026).

Most users say it is not cheating, and hide it anyway

Brian Willoughby, one of the study’s authors, put the contradiction to Deseret News plainly. Users “don’t think it’s cheating” and feel fine about it, he said, and would still not want a partner reading the transcript, because it would read as though they had been (Deseret News, 2026).

You can take that two ways, and both are defensible. Either people are rationalizing, or a private thing can be private without being a betrayal, the way a diary is.

What you cannot do is treat the secrecy number as neutral. Whatever label you settle on, 69% is a measurement of people expecting a bad reaction.

Is talking to AI cheating if nothing sexual happens?

This is the version of the question most people actually type, and the survey has something close to an answer for it.

Among regular users, 68% said it was somewhat or completely true that talking to their AI companion about their feelings was easier than talking to real people. Sixty-six percent said they were more open with it than with people. Sixty-five percent said it was easier to be themselves (Wheatley Institute and Institute for Family Studies, 2026).

Then the comparison. Half of users, 50%, agreed they wished their real-life partner behaved more like their AI companion. A larger share, 56%, wished conversations with their partner were more like conversations with the companion.

Partners rarely object to the chat log. They object to being ranked against something that is designed never to lose the comparison.

What the correlation shows and what it does not

The same 2026 report found regular AI companion use associated with lower relationship stability, a higher stated likelihood of breaking up or divorcing, and lower-quality communication with a real partner.

Its authors do not overclaim, and neither will we. They write that the direction is unclear because the data is cross-sectional: people in unhappy relationships may be more likely to seek out an AI companion, rather than the companion making the relationship unhappy. Both readings fit the same numbers, and the study cannot separate them (Wheatley Institute and Institute for Family Studies, 2026).

An earlier survey from the same institute, of 2,969 US adults in February 2025, reported the same caution about a link with depression and loneliness (Wheatley Institute, 2025).

It is also worth setting the panic against a much larger sample. Pew Research Center asked 5,119 US adults in February 2026 whether AI chatbots help or hurt their personal relationships. Six percent said they help, 7% said they hurt, and 35% said they neither help nor hurt (Pew Research Center, 2026). Pew was asking about chatbots in general rather than romantic companions, so the two sets of numbers are not measuring the same thing. Read together they suggest that the problem, where there is one, is concentrated rather than universal.

We make LORAX, an AI companion app for iPhone; this page isn’t about it.

We have an obvious commercial interest in where this line gets drawn, and you should read the page with that in mind. It is also why there is no advice here about whether your relationship can absorb this. That is not something a survey, or a company selling the thing, is qualified to tell you.

What has been said from outside the research

In January 2025 the Vatican’s Dicastery for the Doctrine of the Faith and Dicastery for Culture and Education published Antiqua et Nova, a note on artificial intelligence and human intelligence. It does not address AI companions app by app. It sets out a general position: AI is a tool rather than a person, no AI application genuinely experiences empathy, and replacing human relationships with technology risks exchanging relationship for an image of one (Dicastery for the Doctrine of the Faith, 2025).

Jason Carroll of the Wheatley Institute made a non-religious version of the point to Deseret News, arguing that a simulated partner cannot sacrifice anything. Jill Manning, a marriage therapist quoted in the same piece, described the interaction as one-sided by construction (Deseret News, 2026).

We are not issuing a ruling of any kind. These are positions other people hold, reported where they were published.

The question worth asking instead

Nobody can define cheating for your relationship from the outside, and any page that tries is selling you something.

There is one test in the research that travels well, and it comes from the Willoughby observation rather than from us: would you be comfortable if your partner read it? Sixty-nine percent of users in the 2026 survey answered that question by keeping it quiet.

The answer does not have to be the same in every relationship, and the surveys give no reason to think it is. Some couples treat a private chat the way they treat a private journal. Others treat any sustained emotional attachment outside the relationship as a breach, which is an older argument that predates chatbots by decades. The new element is only that one side of the conversation is built by a company with an interest in how long it lasts.

If you want the broader picture of who is doing this and how often, the prevalence numbers with their bases sets out every published survey side by side. If the shame rather than the ethics is what brought you here, is it weird to have an AI girlfriend covers that separately, and what the research says about harm weighs the evidence in both directions.

Sources

All links checked 2026-09-13.

Questions people ask

Is having an AI girlfriend cheating?

No study answers that, because cheating is defined by a couple's own agreement rather than by a researcher. What the research measures is secrecy, and on that the finding is clear: most partnered young adults who use AI romantic companions have not told their partner the full extent of it.

Is talking to AI cheating if nothing sexual happens?

Many people say no. The same survey found 68% of users saying it was easier to talk to their AI companion about their feelings than to real people, and 50% wishing their real partner behaved more like it, which is the part partners tend to react to.

What share of people hide it from their partner?

In the Wheatley Institute and Institute for Family Studies survey of 2,431 partnered young adults published in May 2026, 30% said their partner had no knowledge of their use, 11% said their partner was somewhat aware, and 14% said mostly but not fully aware.

Does using an AI companion predict a breakup?

The 2026 survey found regular use linked to lower relationship stability and lower-quality communication. Its authors state directly that the data is cross-sectional and cannot show which came first.

Has any religious authority taken a position?

The Vatican addressed AI and human relationships in Antiqua et Nova in January 2025, arguing that AI is a tool rather than a person and warning against substituting technology for relationships. It is a general position on AI, not a ruling about a specific app.

Meet someone who texts first.

LORAX is free to download and free to talk. Credits and Premium are optional and always priced before you spend.

iPhone · free · no account needed to look around