King's College London

Writing for Frontiers in Science, the team including, Emeritus Professor Prokar Dasgupta – who recently performed the UK’s first long‑distance robotic operation – and Dr Alejandro Granados, suggest that “embodied AI” could safely augment surgical practice and usher in an era of truly personalised surgery.

AI‑enabled robotic systems have the potential to significantly improve surgical teams’ performance, awareness and decision‑making, while also increasing operational effectiveness and efficiency, however, progress will depend on clear regulatory frameworks and sustained human oversight, the paper says.

The analysis highlights several unresolved questions, including how regulators should manage systems that continue to learn and evolve after approval, how bias in datasets can be prevented from reinforcing healthcare inequalities and how global imbalances in research and industry concentration can be addressed.

Professor Dasgupta said the pace of change could be rapid: “Using advanced AI and robotics in the operating room is very exciting. The next few years will see intelligent robots impact all stages of surgery including techniques, emergency responses, team roles, workflows, and assistive functions.”

The paper however, cautions that AI systems must support – rather than disrupt – established surgical practice. Surgeons should remain the primary decision‑makers, supported by strong regulatory and human oversight.

Future advances are expected to include AI embedded directly into surgical robots and connected to sensor‑rich operating theatres. These systems could develop spatial awareness, learn and adapt in real time, benchmark performance, provide autonomous assistance and deliver feedback to teams during procedures.

New large datasets from patients, surgical staff and robotic sensors could also allow AI systems to offer real‑time decision support, helping surgeons optimise each action mid‑operation. Predictive models could also enable surgeons to visualise the likely outcome of different choices before acting, improving patient outcomes through what researchers describe as cause‑and‑effect recognition.

First author Dr Alejandro Granados of King’s College London said surgery was approaching a turning point.

The paper argues that current regulatory approaches, which approve medical devices based on a fixed design however, are poorly-suited to adaptive AI systems that can change after deployment. It suggests reforms to licensing routes, device classification, post‑market surveillance and compliance standards will be necessary to manage higher‑risk technologies.

“AI’s ability to learn presents an unprecedented puzzle. We are at a pivotal time in surgery where we need to begin answering those questions to ensure patients can benefit from the wealth of benefits AI‑powered operating rooms bring,” Dr Granados said.

Standardised clinical trial metrics for AI software, closer collaboration between regulators and professional bodies on training, and new partnerships to ensure lower‑income countries also benefit from innovation are also recommended.

“Human surgeons must continue to be the chief decision‑makers. If we are to maintain the clear chain of authority necessary for safe surgical practice,” Professor Dasgupta said.

Co-author of the paper, Professor Sebastien Ourselin agreed: