How One Man Used AI to Battle His Girlfriend's Recurring Brain Tumor

When his girlfriend's prolactinoma returned for the third time, he stopped waiting for answers from the medical system and started building his own. Armed with nothing but a laptop, a growing sense of desperation, and access to large language model AI tools, he dove into the scientific literature in a way no layperson had realistic access to before.

A prolactinoma is a benign tumor of the pituitary gland that causes overproduction of the hormone prolactin. While rarely life-threatening, it can cause debilitating symptoms including vision disturbances, chronic headaches, infertility, and severe hormonal imbalances. Standard treatment involves dopamine agonist medications, and in resistant cases, surgery or radiation. But for some patients, the tumor keeps coming back.

Frustrated by what he describes as dismissive consultations and a one-size-fits-all treatment approach, he began feeding research papers, clinical trial data, and his girlfriend's medical records into AI systems capable of synthesizing complex information. The goal was not to replace her doctors, but to walk into appointments armed with the right questions — and sometimes, the right answers.

Using tools like ChatGPT and other AI research assistants, he was able to identify patterns in the literature around cabergoline-resistant prolactinomas, surface emerging studies on alternative dosing strategies, and flag potential drug interactions that had not been discussed in her care plan. In one instance, he says an AI-assisted deep dive helped surface a clinical trial she was later enrolled in.

His experience speaks to a broader shift happening quietly in hospitals and living rooms around the world. Patients and their advocates are increasingly using AI not as a replacement for medical expertise, but as a force multiplier — a way to compress years of medical education into targeted, actionable insights when time is critical and trust in institutions is frayed.

Medical professionals have reacted to this trend with a mixture of cautious encouragement and concern. While some physicians welcome engaged, informed patients, others worry about misinformation, misinterpretation of data, and patients making unilateral treatment decisions based on AI outputs that were never designed to serve as clinical guidance.

He is clear that AI did not cure his girlfriend's tumor. But he credits it with giving them agency at a moment when they felt entirely powerless. Her condition has stabilized, and she is currently responding to a revised treatment protocol. Whether AI played a decisive role may be impossible to prove — but for him, the act of fighting back mattered just as much as the outcome.