AI Company Paid Ten People to Masturbate Every Day for a Month. Here Is What the Data Actually Shows.

Key points

• Joi AI, an AI companion platform, paid ten participants $2,000 each to masturbate daily for 28 days in July 2024, using the company's app twice a week.
• Preliminary results from 134 session logs claim stress fell by 25 percent and focus rose by 17 percent, all measured by self-report alone.
• Only six of the ten participants completed the study.
• An independent neuroscientist says the claim that masturbation could help tackle addiction is not supported by the evidence.
• More than 150,000 people applied for the ten spots in the experiment.

Keshav had one new task every morning for twenty-eight days: masturbate, then fill in a log about his mood, energy levels, focus, and whether he was procrastinating or craving alcohol, nicotine, or junk food.

The 38-year-old Indonesian trader (he asked not to be identified by surname for professional reasons) was one of ten paid "masturbation consultants" recruited by Joi AI, a company that sells subscriptions to AI companions: digital characters that users chat and have erotic interactions with. The experiment ran in July. Participants masturbated every day except Sundays, using the Joi AI app twice a week and their own preferred methods the rest of the time. Each received $2,000, as first reported by Wired.

What did the study actually find?

Preliminary results, drawn from 134 self-reported session logs, suggest stress levels dropped by 25 percent and focus climbed by 17 percent. The company also reported that self-reported cravings for nicotine, alcohol, junk food, and doomscrolling fell by 44 percent after participants had masturbated.

Joi AI's head of brand and communications, Julie Levin, called that craving drop "a genuine breakthrough" and suggested masturbation could be "an unexpected way to tackle addictions."

Neuroscientist Nicole Prause, who studies human sexuality and addiction, disagrees sharply. "This is a gross overstatement from the evidence they claim to have," she told Wired. Because masturbation can never have a placebo group (participants always know whether they did it), unbiased self-report is nearly impossible. "Someone saying 'I feel focused' is just that," she said. "There are dozens of cognitive tests that could be used to support a self-report. I do not see any."

Prause is herself running a large laboratory study of male orgasms. She does say masturbation to orgasm is a real method of mood management, and the biggest effects she sees are reduced anxiety and greater relaxation after climax. Mood improvement, though, is a long way from addiction treatment.

Who took part, and how did they feel?

Finding ten participants was not the problem. Over 150,000 people applied. Five were handpicked by the company; five won public competitions, including one judged by porn performers and another that called for the "most unhinged" fetish. The winner had a balloon fetish.

Keshav used the app's video-call feature, where digital companions performed sexual acts while talking. His preferred setting was a virtual airplane bathroom. He reported feeling more clear-headed afterwards and said the experience helped him think about how to approach women in real life.

Sex-tech researcher Chloé Locatelli, based at King's College London, also participated because she is running a separate academic study with Joi AI on how people form relationships with AI characters. She found the app only partially effective for arousal. Her broader concern is one that our earlier piece on China's AI companion boom raised in a different context: what happens when a highly personalised, always-available digital relationship begins to crowd out real ones. As Locatelli put it, "If it becomes a bit of a crutch and your only form of sexual exploration, then that's not something I would necessarily encourage."

What should you make of this?

The science here is thin. Six of ten participants completed the study. Every result rests on self-report, which researchers treat as the weakest form of evidence for anything touching mood or cognition. There is no control group and no independent oversight.

Masturbation does have documented benefits: endorphins and oxytocin, hormones that reduce stress, are released at orgasm, and better sleep is a commonly reported effect. Those are real. The addiction-management claim is not proven, and a company-designed, company-run study with six completers cannot make it so.

Joi AI's app starts at $3.99 a month, with 10,000 in-app coins for premium content costing $99.99. The company says its goal is to "solve male loneliness." That's a genuine problem worth taking seriously. It deserves better evidence than this.

Common questions

Is there any real science behind masturbation improving mood?

Yes, in a limited sense. Orgasm releases endorphins and oxytocin, hormones that lower stress and improve relaxation, and several peer-reviewed studies support this. The much larger claims about reducing addiction cravings are not supported by published evidence.

Is the Joi AI study peer-reviewed?

No. It is an in-house qualitative study: the company designed it, ran it, and analysed the results itself. It has not been reviewed or published in a scientific journal.

Should people be worried about using AI companion apps?

Locatelli's caution is worth taking seriously. AI companions that are highly personalised and always available could, for some users, replace rather than complement real-world intimacy. Occasional use is very different from making them your only form of sexual or emotional connection.