On a whim, Trinity Page Davenport, a 24-year-old office manager from the UK, uploaded a photo of her palm to ChatGPT, asking the AI to read it. She’d heard the tool could offer “palm readings,” breaking down a person’s fate line, heart line, and more. But before it even got to the metaphysical predictions, the chatbot decided to flag something totally unrelated yet concerning.
Rather than something about her love life, the young woman received a stark medical warning that just may have saved her life.
“Before we get into any palm reading,” it said, “I strongly recommend you get that mole on your palm checked out by a dermatologist.”
The mole, it warned, looked like acral lentiginous melanoma — a rare form of skin cancer that tends to appear on the palms, soles of the feet, or under the nails. It’s not the kind of place most people check for cancer, and for that reason, it’s often caught too late.
Came for palm leading, left with a cancer diagnosis
“I expected it to talk about my love life,” Trinity said, “because that’s what I was most interested in finding out about.”
But the response took her by surprise. “It may be nothing,” the AI continued, “but its location and appearance could resemble something called acral lentiginous melanoma… It’s always better to be safe and get a professional evaluation.”
Trinity didn’t panic though, and took the advice to heart. “I was pretty calm,” she said. “I have had a lot of health issues in the past and it’s just something else I need to get looked at.”
She had first noticed the mole in 2020, shortly after the birth of her son. “I never considered it to be anything dangerous,” she said. Now, she’s waiting for a formal dermatological appointment.
Though no diagnosis has been made yet, Trinity’s story quickly went viral in a video that’s been viewed over 4.4 million times.
“Put a finger down,” she began, “if you asked AI to do a palm reading on you, and they told you to get your mole checked because it looks like acral lentiginous melanoma.”
From Fortune to Flesh
Acral lentiginous melanoma, the kind ChatGPT flagged, is especially easy to miss. It often develops in areas not exposed to sunlight, which means people don’t think to look for danger there. It also disproportionately affects people with darker skin tones, making early detection a persistent public health challenge.
ChatGPT, developed by the San Francisco-based company OpenAI, is better known for writing essays, translating languages, or summarizing news. But as people experiment with its visual capabilities — uploading photos to generate artwork, identify objects, or even read palms — sometimes astonishing uses are emerging.
“I will be really grateful if this is an instance in which a robot saved my life,” Trinity said. “And I hope that more people use the technology if they’re concerned.”
AI is actually extremely good at spotting cancer. Researchers have been developing machine learning models for years to help dermatologists identify melanoma. In peer-reviewed trials, some AIs have performed as well as — or better than — human specialists at spotting certain types of skin cancer.
But there’s a crucial difference: those tools were trained on curated medical datasets and deployed in clinical settings. ChatGPT wasn’t. It’s not certified for medical diagnostics, nor was it ever intended to act as a digital dermatologist.
The Age of Digital Health
Even OpenAI has issued reminders that its chatbot should not be used for health decisions. And NHS England agrees. According to its guidance, any mole that changes size, shape, or color — or becomes itchy or painful — should be checked by a professional.
Still, this story strikes a chord in the age of digital health. As one commenter on TikTok wrote, “Me staring at a mole on my palm that looks VERY similar.” Another shared, “Melanoma survivor here. I always remember the doctor telling me the worst spot to have melanoma is the palms of your hands or bottom of your feet.”
For Trinity, the experience has already reshaped her view of AI. “It’s really cool to think about the ways that AI can help people,” she said. “If this is something that actually saves my life, then yeah—I’m all for it.”
The irony isn’t lost on anyone. She had approached the world’s most advanced AI for a reading rooted in pseudoscience — and it responded with a message grounded in the logic of modern dermatology.
