London surgeons used AI to remove a brain tumour, and their patient kept his sight

Key points

• Surgeons at the National Hospital for Neurology and Neurosurgery in London completed the world's first AI-assisted brain tumour removal in May 2025.
• Patient Rhys Hibbert, 48, kept his sight after the procedure targeted anatomy that could have blinded him.
• The AI system read live footage from the surgical camera to identify structures the team needed to avoid.
• Details were withheld for several weeks while Hibbert recovered, then released on Thursday.

Rhys Hibbert went into surgery knowing what was at risk. The tumour sat close to structures that govern vision. One wrong cut and he'd likely never see normally again.

He came out with his sight intact.

Surgeons at the National Hospital for Neurology and Neurosurgery (NHNN), part of University College London Hospitals NHS trust, performed the operation in May 2025. What set it apart from every prior brain tumour removal was a layer of real-time AI assistance: software watching the live camera feed from inside the skull and highlighting critical anatomy as the team worked.

Think of it as a satnav overlay for surgery. The operating team could see the tumour, but the AI continuously flagged surrounding structures that carry signals between eye and brain, tissue that can look almost identical to what needs removing.

What did the AI actually do?

Frame by frame, the system identified anatomical landmarks and alerted the surgeons to high-risk areas. It didn't hold a scalpel or make decisions. Think of it as a second pair of eyes that never blinks and doesn't tire.

That matters to anyone who's wondered whether a surgeon is still sharp deep into a long shift. Fatigue is real, attention drifts, and a persistent AI alert saying "don't cut here" adds a genuine safety net.

Does this mean AI is now doing surgery?

No. The surgeons were in control throughout. The AI supplied information; the humans acted on it. That distinction matters, and it's worth keeping in mind as the coverage spreads.

The NHNN is among the most specialised neuroscience centres anywhere, and this was a carefully controlled first case, not a rollout. The Guardian reported it first.

A successful debut in brain surgery carries real weight, though. The brain offers almost no margin for error. If AI-assisted visual analysis can work there, it opens a credible path toward similar tools in spinal surgery, eye operations and reconstructive work, where the cost of a slip is permanent.

Our earlier story on the AI scribe now handling 2.7 million patient visits a week showed how quickly clinical AI can scale once trust is established. The NHNN case is an earlier, harder problem solved.

Hibbert spent weeks recovering before the hospital released any details. The fact that he's well enough for the story to be told is, on its own, good news.

Common questions

Is AI-assisted surgery available at other hospitals?

Not yet in this form. This was a world first. The technology will need further trials and regulatory review before other hospitals can adopt it routinely.

Should patients ask about AI tools before their own operations?

Asking your surgical team what decision-support tools they use is always reasonable. This case shows AI assistance can be a safety benefit rather than something to fear, but outside specialist centres the technology is still early-stage.