World's first AI-assisted brain surgery successfully removes patient's tumour

Rhys Hibbert was walking to lunch one day when his body stopped obeying him. He collapsed on the road and began seizing. A brain scan showed why: an 11-millimetre tumour growing on his pituitary gland, a marble-sized organ at the base of the skull that regulates hormones controlling growth, metabolism, blood pressure, and temperature. The tumour itself was benign, but its location was treacherous. It sat inches from the carotid arteries that feed blood to the brain and the optic nerves that control vision. As it swelled, Hibbert's peripheral vision began to narrow. He tripped over things. Fatigue and dizziness became his constant companions. For 18 months, the 48-year-old customer services manager from Bedfordshire watched his life contract.

When surgeons at University College London Hospital offered him a choice—conventional surgery or something untested—Hibbert did not hesitate. He would be the first person on Earth to have brain surgery guided by live artificial intelligence. "If patients are not prepared to join research how can doctors ever learn and how can medicine ever progress?" he said. The conventional approach carried real stakes. Surgeons typically perform 10 to 20 of these pituitary tumour removals annually, relying on pre-operative scans to memorise each patient's unique anatomy. Even with that preparation, there was a 25 to 50 percent chance they would not remove the entire tumour, and a 0.5 to 2 percent chance they would rupture a major blood vessel—a catastrophic outcome.

The surgical team, led by Professor Hani Marcus, had spent years training an AI system on hundreds of recorded pituitary operations. They had painstakingly drawn around blood vessels and nerves in each video, teaching the algorithm to recognise the hidden anatomical landmarks that separate safe removal from disaster. The technology worked like facial recognition on a smartphone, except instead of identifying faces, it identified the critical structures surgeons needed to avoid. On the day of Hibbert's operation, surgeons inserted an endoscope—a thin camera on a stick—through his nose to the base of his skull. As they worked, a second screen displayed the live video feed analysed in real time by the AI system. It tracked the surgical instruments, marked the locations of vessels and nerves, and highlighted the safest zones for tumour extraction. The AI had been trained on more operations than most surgeons perform in a lifetime, and it could offer guidance without the fatigue that accumulates during hours of intense concentration.

The surgery succeeded. When Hibbert opened his eyes, the world had changed. "It felt like I had got 360-degree vision," he said. "I hadn't seen that clearly for a year." In the eight weeks that followed, his sight improved every few days. The fatigue vanished. The dizziness disappeared. The emotional weight of being a burden—of not being able to make his wife a cup of coffee in the morning as he had for 18 years—lifted. "It's given me my life back," he said.

The research team, funded by the National Institute for Health and Care Research and Google, emphasised that the AI was an assistant, not an autonomous surgeon. The human operators retained full control and could override the system's recommendations. Their long-term vision is more ambitious: to develop what they call a surgical ChatGPT, an expert system that sits in the operating room as a second opinion, available for consultation but never in command. Larger trials are now underway. The implications extend far beyond pituitary tumours. Any surgery involving hidden critical structures—vessels, nerves, vital organs—could potentially benefit from real-time AI guidance. For now, Hibbert's case stands as proof that the technology works, and that patients willing to pioneer new approaches can help reshape what medicine is capable of doing.