First real-time AI use in brain surgery protects patient's sight

Overview

Updated Sep 7

Rhys Hibbert, 48, went into surgery with an 11mm tumour pressing on the nerves that control his vision. He came out with what he calls 360-degree panoramic sight.

The difference: an AI system that watched the live surgical video feed and marked the blood vessels and nerves the surgeon needed to avoid. The May operation at the National Hospital for Neurology and Neurosurgery in London was the first time AI guided a neurosurgeon in real time during live surgery.

The system, developed at University College London, analyzed the camera feed from an endoscope inserted through the nose. It tracked surgical instruments and marked risky areas near the optic nerves and carotid arteries. A larger clinical trial is now planned.

Why it matters

If this AI proves out in larger trials, delicate brain surgery could get safer for thousands of patients a year.

Questions about this story

Free account needed to ask — your question is kept and asked for you right after sign-up. Answers are public.

No questions yet — be the first to ask.

Key Indicators

11mm

Tumour size

Non-cancerous growth on the pituitary gland, close to optic nerves and carotid arteries.

25-50%

Chance of incomplete removal without AI

Standard complication rate for this procedure without AI assistance.

0.5-2%

Risk of major blood vessel injury

Risk of injuring a carotid artery during standard surgery.

Patient age

Rhys Hibbert was slowly going blind before the operation.

Voices

Curated perspectives — historical figures and your fellow readers.

People Involved

Rhys Hibbert

First patient to undergo live AI-assisted brain surgery

Recovered, sight improved dramatically

Professor Hani Marcus

Consultant neurosurgeon at NHNN, UCL Queen Square Institute of Neurology

Performed the first AI-assisted surgery

Danyal Khan

UCL PhD candidate and UCL/UCLH neurosurgical resident

Co-led the first AI-assisted surgery

Dr Sophia Bano

Associate Professor in Robotics and AI at UCL Computer Science and UCL Hawkes Institute

Technical lead for the AI system

Organizations Involved

University College London (UCL)

University and research institution

Developed the AI system at the UCL Hawkes Institute

Developed the real-time AI system used in the first live AI-assisted brain surgery.

University College London Hospitals NHS Foundation Trust

NHS hospital trust

Hosted the surgery at the National Hospital for Neurology and Neurosurgery

The NHS trust where the first live AI-assisted brain surgery was performed.

National Institute for Health and Care Research

Government Research Funder

Co-funded the clinical trial with Google

UK government body that co-funded the clinical trial of the AI-assisted surgery system.

Google

Technology company

Co-funded the clinical trial with NIHR

Co-funded the clinical trial of the AI-assisted surgery system.

Timeline

May 2026 September 2026

4 events Latest: September 4th, 2026 · 2 months ago

Tap a bar to jump to that date

September 2026
Human Progress reports on the milestone
Latest Media

Human Progress publishes its report on the first real-time AI use in brain surgery.
September 4th, 2026
August 2026
Global media report the milestone
Media

Guardian, Independent, Euronews and other outlets cover the world-first procedure.
August 27th, 2026
UCL and UCLH announce the surgery
Announcement

Details of the first live AI-assisted surgery revealed after the patient recovered.
August 26th, 2026
May 2026
First live AI-assisted brain surgery performed
Medical

Rhys Hibbert's 11mm pituitary tumour removed with real-time AI guidance at NHNN.
May 2026

Scenarios

Larger trial confirms AI-assisted surgery is safe and effective

Likely Resolves by End of 2027

Discussed by: UCL and UCLH researchers, NIHR

The ongoing clinical trial reports positive feasibility, safety, and clinical outcomes. UCL and UCLH expand the trial to more patients, and AI-assisted pituitary surgery becomes the standard approach at NHNN.

AI expands to other neurosurgical procedures

Possible Resolves by Q2 2028

Discussed by: Professor Hani Marcus, UCL researchers

The technology is adapted for other delicate procedures near critical structures. UCL announces new clinical trials for different types of brain surgery, moving toward Marcus's vision of a 'ChatGPT for surgeons.'

Trial results are mixed, AI stays a training tool

Unlikely Resolves by End of 2027

Discussed by: Not explicitly predicted by sources; represents the conservative outcome if the trial shows limited benefit

The larger trial shows the AI provides limited benefit in real surgery, or surgeons find the live overlays distracting. The technology continues to be used for training but doesn't become standard surgical practice.

Historical Context

3 moments from history that rhyme with this story — and how they unfolded.

1990s

Image-guided neurosurgery (1990s)

Neurosurgeons began using frameless stereotactic navigation systems that tracked surgical instruments against pre-operative MRI and CT scans. This replaced reliance on anatomical landmarks and X-rays during surgery.

Then

Brain surgery became more precise, with better tumour resection rates and fewer complications.

Now

Image-guided navigation is now standard in neurosurgery worldwide.

Why this matters now

The current AI system builds on image-guided navigation but adds live analysis of the surgical video feed, rather than relying on pre-operative scans.

2000

da Vinci Surgical System (2000)

The FDA approved the da Vinci system, the first robotic-assisted surgical platform. It took years of clinical trials, surgeon training, and cost negotiations before robotic surgery became widespread.

Then

Robotic surgery was initially limited to a few procedures at major hospitals.

Now

da Vinci is now used in millions of procedures annually worldwide.

Why this matters now

The adoption curve for AI-assisted surgery may follow a similar path, starting with a single procedure and expanding as evidence accumulates.

2011

IBM Watson in medicine (2011)

IBM promoted Watson as a breakthrough in AI-assisted medicine, particularly for oncology treatment recommendations. It was deployed at several hospitals with high expectations.

Then

Watson's recommendations were often inconsistent with clinical practice and it struggled to integrate into hospital workflows.

Now

IBM sold Watson Health in 2022, a cautionary tale about AI hype in medicine.

Why this matters now

The UCL team emphasizes that the AI is an assistant, not a replacement, and that surgeons retain full control. Clinical validation will determine whether this AI avoids Watson's fate.