By
Stuart Kerr, Technology Correspondent, LiveAIWire

A surgical robot completed a segment of bowel reconnection without
any human guidance during a 2022 trial at Johns Hopkins University — the
first fully autonomous soft-tissue surgery demonstrated in a live animal
model. The procedure was more consistent and produced fewer complications
than the same procedure performed by experienced surgeons. The researchers
were careful to note that full clinical autonomy remained years away. The
direction of travel was not in doubt.

AI is entering the operating room from multiple angles
simultaneously: as a guidance system for surgeons, as an autonomous operator
for specific procedural steps, as a real-time monitor for complications, and
as a planning tool that designs the surgical approach before the patient is
anaesthetised. The cumulative effect is a transformation of surgical practice
that is already under way in major hospital systems and will reach the
mainstream of surgical care within the decade.

Robotic Surgery: From Assistance to Autonomy

The da Vinci surgical system, manufactured by Intuitive Surgical,
has been in clinical use since 2000 and has performed millions of procedures.
It is not an autonomous system: the surgeon controls every movement through a
console, and the robot translates those movements into the precise,
tremor-free actions of instruments too small for unassisted human hands. The
AI component provides stability, scaling, and visualisation enhancement, but
the decision-making is entirely human.

The current generation of surgical AI is moving beyond this purely
assistive model. Systems are being developed that can identify the optimal
incision path, recognise anatomical landmarks and flag proximity to critical
structures, distinguish healthy tissue from diseased tissue in real time
during a procedure, and alert the surgeon when a planned action deviates from
the pre-operative plan in ways that increase complication risk. These are
guidance functions, not autonomous actions, but they represent a meaningful
shift in where surgical intelligence resides.

Research published in Nature
Medicine has documented AI systems that outperformed specialist
surgeons in identifying cancerous tissue margins during surgery, reducing the
rate of incomplete resections that require follow-up operations. The clinical
benefit is concrete: fewer re-operations, better patient outcomes, lower
system costs.

Planning Before the First Incision

AI’s contribution to surgery begins before the patient enters the
operating room. Preoperative planning systems trained on imaging data can
analyse CT and MRI scans to generate three-dimensional models of the surgical
site, identify anomalies in vasculature or anatomy that would complicate a
standard approach, and recommend patient-specific procedural strategies.
Surgeons using these systems report that pre-operative AI planning surfaces
considerations they would not have identified from conventional imaging
review alone.

In orthopaedic surgery, AI planning tools have demonstrated
particular value. Systems that model bone geometry, implant positioning, and
biomechanical load distribution preoperatively have been associated with
improved implant alignment and reduced revision rates in knee and hip
replacement surgery. The evidence base here is now substantial enough that AI
planning is becoming standard practice in high-volume orthopaedic
centres.

What this means for you as a patient: if you are having elective
surgery at a major hospital centre, AI is almost certainly involved in your
care, whether or not it has been explicitly disclosed. Knowing what tools are
in use and what role they play in your surgeon’s decision-making is a
reasonable question to ask during your pre-operative
consultation.

Real-Time Monitoring and Complication Detection

Intraoperative AI monitoring systems analyse physiological data
streams in real time, identifying early indicators of complications that
might not be immediately apparent to the surgical team. Haemodynamic
instability, early signs of surgical site bleeding, anaesthetic depth
anomalies, and tissue perfusion changes can all be detected algorithmically
before they manifest as clinical emergencies.

These monitoring applications address one of the most consistent
findings in surgical safety research: complications that are identified and
managed early have dramatically better outcomes than the same complications
identified late. The World
Health Organization’s patient safety data consistently identifies
intraoperative complications as a leading cause of preventable surgical
mortality, and AI monitoring represents a meaningful intervention point for
reducing that burden.

The Autonomy Frontier and Its Governance

The Johns Hopkins bowel reconnection trial points toward a future
in which AI systems perform defined procedural steps autonomously, with a
supervising surgeon available to intervene rather than controlling every
movement. This model, described as supervisory control, mirrors the
architecture of autonomous vehicle systems, where the AI handles routine
operation and the human maintains oversight for edge cases and
exceptions.

The regulatory and liability framework for autonomous surgical
steps does not yet exist in any jurisdiction. If an AI system performing an
autonomous procedural step makes an error, who is responsible: the surgeon
who approved the use of the system, the hospital that deployed it, or the
manufacturer that trained and certified it? These questions are not
hypothetical as the technology approaches clinical deployment, and the answers
will shape how quickly autonomous surgical AI is adopted and in which
healthcare systems.

The fundamental
challenge of trusting AI in high-stakes medical decisions is
sharpest in surgery, where the consequences of an error are immediate,
physical, and potentially irreversible. The evidence base for surgical AI is
growing rapidly; the governance infrastructure to deploy it responsibly is
developing more slowly.

Access, Equity, and the Surgical AI Divide

Robotic surgical systems are expensive to purchase, maintain, and
staff. The da Vinci system costs in the range of one to two million dollars
per installation, with significant ongoing maintenance and consumable costs.
This investment is viable for large hospital systems with high procedure
volumes; it is not viable for district hospitals, rural health services, or
the healthcare systems of most low- and middle-income
countries.

The result is a surgical AI divide that mirrors the digital divide
in other domains: the patients who benefit most readily from AI-enhanced
surgery are those in well-resourced urban hospital centres, while patients in
under-resourced settings continue to receive care without the technology.
Addressing this divide requires either significant cost reduction as the
technology matures or deliberate policy intervention to ensure that surgical
AI investment is distributed according to health need rather than purchasing
power.

The broader pattern of AI
tools delivering unequal benefits across differently resourced
populations is consistent across healthcare and beyond. In surgery,
where the intervention is irreversible, the stakes of that inequality are
unusually high.

The governance questions around autonomous surgical AI mirror
those emerging in other
domains where algorithms make consequential decisions without
transparent reasoning — liability and accountability frameworks are
consistently underdeveloped relative to the pace of
deployment.

Training implications are significant for
surgical education. If AI systems increasingly handle the most technically
demanding aspects of procedures, trainees may receive less exposure to those
elements. Several surgical training programmes have begun addressing this
explicitly, designing curricula that ensure trainees develop foundational
technical competence independently of AI assistance before working with
AI-augmented systems.

The pace of AI adoption in surgery is
also reshaping the economics of the medical device industry. Companies that
have invested in AI surgical guidance capabilities are capturing market share
from conventional device manufacturers, and the competitive pressure is
accelerating investment across the sector. For patients, the consequence is
that AI-enhanced surgical tools are likely to become standard rather than
premium over the next decade, normalising the human-AI collaboration model in
operating rooms that currently only the best-resourced hospital systems
deploy at scale.

About the Author

Stuart
Kerr is a technology correspondent at LiveAIWire, covering artificial
intelligence, emerging technologies, and their impact on society and
industry.