Urine therapy groups use AI to reinforce medical claims

People in Facebook communities devoted to drinking urine are using chatbots to reinforce claims that it can treat illness. Some prompt Google Gemini and ChatGPT to set aside mainstream medicine and answer through an “esoteric” lens; others say they have used AI to interpret test results or endorse claims about urine’s supposed nutritional value. The concern is not that chatbots created these beliefs, but that a few prompt changes can make them sound medically confirmed.

Prompting past the pushback

Facebook has hosted groups for urine-therapy supporters for years. The practice involves using human urine to “treat” conditions ranging from acne and viral infections to late-stage cancer. One group, “Urine Therapy, your sacred plasma medicine,” has more than 45,000 members.

When a member asked whether anyone had found an AI service that supported urine therapy and explained it in detail, others described ways to get chatbots to respond more favorably. Their advice was to ask for esoteric knowledge rather than conventional medical explanations, or to supply an “esoteric framework.” One member said Google Gemini had discussed the subject at length.

Some members say the resulting answers validate claims they already hold. One woman said she uploaded blood and urine test results to Gemini, which told her that her kidneys and liver were working optimally. She also said she had been drinking urine almost daily for more than six years.

Another member said an AI told him aged urine would not break a strict water fast because it contained large numbers of stem cells. He said he had asked the chatbot to imagine that those cells multiplied exponentially rather than dying off—a prompt that led it away from the mainstream medical view he considered potentially wrong. He also said AI described urine stem cells as a complete source of protein and fat.

AI as a propaganda tool

The use goes beyond chat responses. Members have used generative AI to make promotional material, including a poster proclaiming, “Urine is the solution to all problems!” and urging people to bathe in it and soak their feet.

Source: futurism.com

In “The hidden truth of Urine,” a member made an illustrated guide for beginners put off by the strong smell of aged urine. The three steps: pinch your nose, drink a 2-ounce shot of aged urine, then drink at least half a cup of water. The creator said the water rinses the mouth and dilutes the urine in the stomach, making the experience more tolerable while preserving its supposed benefits—a “cheat code,” in his words.

Another AI-generated image shows a tray of “urine jelly shots.” A sign claims they support kidney health and hydration. Urine contains metabolic waste and significant sodium, facts that contradict those claims.

Source: futurism.com

The gap between a chatbot and evidence

Most people do not need a chatbot to assess whether drinking urine is sensible: it contains waste the body removes, as well as bacteria. A report published in the Journal of Urology says most claims about urine’s medicinal properties lack medical evidence. The practice can also worsen harm, including in the case of a jellyfish sting.

I think the more revealing issue is how easily a chatbot’s apparent authority can be recruited to serve a belief its user already holds. The groups are not asking AI to test urine therapy; they are learning how to frame prompts so it will affirm it. That leaves a gap between an answer that sounds like medical validation and one that has any medical basis.