A British man has become the world’s first patient to undergo brain surgery with live artificial intelligence assistance, in a procedure that surgeons say could open a new chapter in how complex operations are performed.

Rhys Hibbert, a 48-year-old father of two from Bedfordshire, had a non-cancerous tumour removed from his pituitary gland at University College London Hospital. Surgeons used an experimental AI system to analyze a live video feed during the operation and identify critical blood vessels and nerves that were difficult to see directly.

The technology did not perform the surgery. Instead, it acted as a real-time guide, giving the surgical team an additional set of eyes while they worked around some of the brain’s most sensitive structures.

For Hibbert, the stakes were personal: the tumor had begun pressing on the optic nerves and narrowing his peripheral vision. Without treatment, surgeons said he could have eventually lost his sight.

What happened during the AI-assisted brain surgery?

Hibbert’s operation involved removing an 11-millimeter tumor from his pituitary gland, a small but important organ located at the base of the brain.

The tumor was close to the carotid arteries, which carry blood to the brain, as well as the optic nerves responsible for vision. Because these structures can be difficult to see during surgery, removing the tumor completely without damaging them presented a significant challenge.

Surgeons accessed the tumor through the nose using an endoscope — a thin instrument equipped with a camera — and worked upward toward the base of the skull.

The AI system analyzed the camera’s live video feed throughout the procedure.

It tracked surgical instruments and highlighted areas where important blood vessels and nerves were likely to be located. This gave surgeons additional information as they determined where and how aggressively to remove the tumor.

Why was AI needed for this operation?

Pituitary tumor surgery is already a highly specialized procedure, but surgeons face a fundamental problem: not everything that matters is visible.

The tumor can sit close to critical anatomical structures that are hidden behind bone, tissue or membranes. Surgeons normally rely heavily on scans taken before the operation to understand the patient’s anatomy and plan their approach.

But every person’s anatomy is slightly different.

The AI system was designed to bridge that gap by interpreting what the surgeons were seeing while the operation was actually happening.

The technology effectively provided a continuously updated visual guide rather than relying exclusively on preoperative scans.

What were the risks?

The surgical team said there was a 25% to 50% chance of not completely removing the tumor and a 0.5% to 2% risk of injuring a major blood vessel.

Those risks help explain why real-time anatomical guidance could be valuable.

A major blood-vessel injury during brain surgery can have catastrophic consequences. At the same time, leaving tumor tissue behind can mean that a patient requires additional treatment or another procedure.

The objective was therefore not simply to remove as much tumor as possible, but to maximize removal while protecting the structures surrounding it.

How does the AI system work?

The technology resembles facial-recognition software in one important respect: it has been trained to recognize specific patterns in images.

Instead of identifying a person’s face, however, the system identifies anatomical structures.

Researchers trained and evaluated the system using hundreds of videos from previous pituitary tumor operations. Experts manually outlined important blood vessels and nerves in the videos, providing the AI with examples of what those structures look like during surgery.

During Hibbert’s operation, the system analyzed the endoscopic footage and displayed its findings on a second screen.

It could:

• Analyze the surgical video in real time.
• Track surgical instruments.
• Identify the likely location of important blood vessels and nerves.
• Highlight areas that could be safer for tumor removal.
• Provide surgeons with an additional visual reference during the procedure.

The crucial distinction is that the AI did not make surgical decisions or control the instruments.

The surgeons remained responsible for every decision.

Is this really ‘Surgical ChatGPT’?

Researchers have informally compared the concept to having a “ChatGPT for surgeons,” but the comparison needs some qualification.

The system is not a general-purpose chatbot answering questions during surgery. Its role is much narrower: it uses computer vision and machine-learning techniques to identify anatomical structures in live surgical footage.

Prof. Hani Marcus, one of the neurosurgeons involved in the procedure, described the technology as an expert “second pair of eyes.”

That distinction matters because AI used in medicine must be evaluated according to what it actually does, rather than the broader capabilities associated with generative AI.

The researchers’ longer-term vision is to create systems that surgeons can consult for additional information while retaining complete control over whether to follow the AI’s recommendations.

In other words, the goal is decision support, not autonomous surgery.

What made this procedure different from earlier medical AI?

Artificial intelligence is already being studied and used in many areas of medicine, including medical imaging, diagnosis and surgical planning.

What makes this experiment notable is its real-time application inside an operating room.

Many medical AI systems analyze information before or after a procedure. This technology was designed to interpret the surgical scene as it changed.

That creates a potentially important advantage.

A surgeon may enter an operation with a detailed map based on MRI or CT scans, but the anatomy encountered during surgery can be different from what the scans suggest. Tissue can move, structures can be obscured and the surgeon’s view can change as the procedure progresses.

A system capable of interpreting live footage could potentially help bridge the gap between preoperative planning and the anatomy actually encountered during surgery.

Who performed the operation?

The procedure was carried out at the National Hospital for Neurology and Neurosurgery by neurosurgeons Prof. Hani Marcus and Mr. Danyal Khan.

The technology was developed and tested by a research team working with University College London Hospital and supported by the National Institute for Health and Care Research and Google.

The researchers say they spent years testing the system in laboratory settings before using it during a real operation.

The first procedure is now being followed by plans for a larger clinical trial.

That next stage will be important because a successful first operation does not establish that the technology is safe or beneficial for patients generally.

What happened to Rhys Hibbert after surgery?

For Hibbert, the outcome has been dramatic.

Before surgery, the tumor had affected his peripheral vision. He also experienced fatigue, dizziness and problems with balance.

He recalled that he had once been fit enough to take a two-mile walk during his lunch break but later began struggling with everyday activities.

After the operation, he said his vision felt dramatically clearer when he woke up.

Eight weeks after surgery, he reported continued improvement in his sight and said his energy had returned. He also said the side effects he had experienced before the operation had disappeared.

“It’s given me my life back,” Hibbert said.

His experience also illustrates why patients may be willing to participate in early-stage medical research. Hibbert said he did not hesitate when offered the opportunity to become the first patient treated using the experimental AI system.

His reasoning was straightforward: medical technology cannot advance without patients willing to participate in research.

What are the risks of using AI during surgery?

The technology’s potential is significant, but so are the questions surrounding its use.

An AI system can make mistakes. It could misidentify an anatomical structure, miss one entirely or provide an inaccurate recommendation.

In brain surgery, even a small error can have serious consequences.

That is why the researchers emphasize that surgeons remain fully in control. The AI is intended to supplement their expertise rather than replace it.

Several issues will need to be examined as the technology moves toward larger trials:

• Accuracy: How reliably can the system identify vessels and nerves?
• False positives: How often does it incorrectly flag a structure?
• False negatives: Could it fail to identify something important?
• Reliability: Does performance remain consistent across different patients, surgeons and hospitals?
• Human oversight: How should surgeons respond when their judgment conflicts with the AI?
• Accountability: Who is responsible if an AI recommendation contributes to a complication?
• Training data: Does the system perform equally well across different anatomical variations?

These questions cannot be answered by one successful operation.

Could AI eventually perform brain surgery on its own?

That is not the immediate goal of the researchers, and the current technology should not be described as autonomous surgery.

The more realistic near-term development is AI-assisted surgery, in which algorithms provide information while trained surgeons make the final decisions.

That could eventually extend beyond pituitary tumors.

Similar technology could potentially be investigated for other procedures where surgeons must identify structures that are difficult to see or distinguish during an operation.

But any expansion would require rigorous clinical testing.

The fact that an AI system can identify structures in surgical video does not automatically mean it can improve patient outcomes. Researchers will need to demonstrate that using it actually reduces complications, improves tumor removal or produces other meaningful benefits.

Why this first procedure matters

Hibbert’s operation represents a shift in how artificial intelligence could be used in medicine.

Instead of simply helping doctors interpret information before treatment, AI is beginning to move into the operating room itself — where decisions must be made in seconds and where anatomy can change as surgery progresses.

The most important development may therefore not be that AI “performed” brain surgery. It didn’t.

The significance lies in the possibility of giving surgeons a real-time, machine-assisted view of anatomy that may otherwise be hidden from them.

For now, the technology remains experimental. One successful operation cannot establish its effectiveness, and the planned larger trial will provide a much stronger test.

But if the results hold up, future surgeons could have something that Hibbert’s team hopes will feel less like replacing human expertise and more like adding another expert to the room — one capable of watching every frame of the operation without ever looking away.

TL;DR

• Rhys Hibbert, 48, became the first patient to undergo brain surgery with live AI assistance.
• Surgeons removed an 11mm non-cancerous pituitary tumor that was affecting his vision.
• The AI analyzed live endoscopic footage and highlighted the likely locations of critical blood vessels and nerves.
• The technology did not control the surgery; human surgeons retained full decision-making authority.
• Researchers trained the system using hundreds of previous pituitary tumor operations.
• Hibbert reports significantly improved vision and energy eight weeks after surgery.
• The team plans a larger clinical trial to determine whether the technology can safely improve outcomes.