More on Autonomous Robotic Surgery

When I am sitting at the robotic console during a cystectomy, the relationship between my hands and the instruments is absolute. Every movement—scaled, filtered, stabilized—originates from me. The robot does nothing independently. This is teleoperation, the model that has defined robotic surgery for more than twenty years.

Despite this reality, headlines and academic discussions are increasingly filled with claims about “autonomous” surgery. For surgeons actually operating, it is important to separate enthusiasm from what is technically and clinically achievable. Urology has led the adoption of robotics, but we are not approaching surgeon replacement. What we are seeing instead is incremental delegation of narrowly defined tasks. The shift is not from surgeon to machine, but from passive tool to intelligent assistant, with the surgeon remaining the legally and clinically responsible decision-maker.

Teleoperation, Automation, and Autonomy Are Not the Same

Much of the confusion stems from imprecise use of language. Automation and autonomy are often treated as interchangeable. They are not.

Teleoperation remains our current standard. Systems like the da Vinci initiate no action on their own and stop immediately when the surgeon disengages. Automation refers to scripted, deterministic execution in controlled environments. Urology has seen this before. The PROBOT performed TURP using a predefined plan, but it could not adapt to bleeding, tissue shift, or deformation. True autonomy requires real-time sensing, interpretation, and adaptation within an unstructured biological environment.

This distinction exposes the common “aviation analogy” as flawed. Commercial aviation operates in a highly structured, physics-governed environment. Surgery does not. Soft tissue deforms unpredictably, vascular anatomy varies at the millimeter scale, and the operative field fundamentally changes with each action. Surgical autonomy is not autopilot.

Where We Actually Are: The LASR Framework

The Levels of Autonomy in Surgical Robotics (LASR) framework is a useful way to ground this discussion.

At Level 0, the surgeon has continuous control. Level 1 systems assist with information or constraints—augmented reality overlays during partial nephrectomy are a good example. These tools guide the surgeon but do not act.

Level 2 introduces task autonomy. MRI-guided prostate biopsy platforms can autonomously align needle guides to coordinates, but the surgeon still performs the biopsy itself.

Level 3 is where meaningful autonomy exists today. Aquablation is the clearest example. The surgeon defines treatment boundaries and safety zones; the system executes the ablation. Outcomes are more consistent and less dependent on surgeon volume.

Levels 4 and 5 remain theoretical. A robot that independently performs a vesicourethral anastomosis or an entire operation without supervision faces unresolved technical, ethical, and liability barriers.

What Is Actually Working

Aquablation succeeds because the task is bounded, the anatomy is constrained, and the system operates under supervision. It standardizes a procedure that historically showed wide variability in outcomes.

Research platforms like the Smart Tissue Autonomous Robot (STAR) show what may eventually be possible for soft-tissue reconstruction. STAR has demonstrated superior consistency in intestinal suturing compared with human surgeons, using real-time visual tracking of tissue deformation. Leak pressures were substantially higher than with manual laparoscopy.

That said, the gap between laboratory success and operative reality remains large. STAR is slow, constrained by safety-driven speed limits and motor precision. These limitations matter in real cases where time, bleeding, and fatigue are not abstract variables.

Other systems highlight where machines already outperform humans. Autonomous navigation platforms in ureteroscopy improve spatial coverage and reduce cognitive workload. These are quiet successes, not headline-grabbing revolutions.

Why Urology Is Both Ideal and Difficult for Autonomy

Urology presents a paradox. We operate on tubular structures and targets that are often well-imaged preoperatively. At the same time, we work in highly deformable soft tissue. The moment a space is entered or an organ mobilized, static imaging becomes unreliable.

Registration drift is a major unsolved problem. For example, in prostate biopsy, the prostate can deform several millimeters with probe insertion alone. Without real-time tracking, autonomous systems risk acting on outdated maps.

The Surgeon’s Role Going Forward

The future is augmentation, not displacement. The most meaningful benefit may be performance standardization. Intelligent systems can enforce safety boundaries, identify optimal motion patterns, and reduce variability between surgeons.

At the same time, there is a real risk of deskilling. Delegating repetitive tasks like suturing raises questions about how trainees will maintain the ability to take over when systems fail. Even current robotic platforms demonstrate steep learning curves, as reflected in longer operative times when new trainees rotate in.

Training will need to adapt. Manual proficiency and emergency takeover must be explicitly taught and assessed. Judgment remains human. Managing an unexpected hemorrhage or abandoning a planned reconstruction requires contextual reasoning that current systems cannot replicate. Yet.

A Realistic Timeline

Regulatory oversight will remain conservative. All current platforms are classified as surgeon-controlled devices, and that will not change quickly.

In the near term, expect more assistive intelligence: real-time structure identification, virtual boundaries, and intraoperative decision support. In the medium term, we may see task-bounded autonomy for discrete phases of operations under supervision.

As I have said before, what we will not see soon is a fully autonomous surgeon.

Outcomes First

Innovation is not the endpoint. Continence, potency, complication rates, and oncologic control matter more than novelty. Robotics will continue to evolve, and the cockpit will become more intelligent. But the surgeon remains the pilot. A pilot guiding the journey through a foggy swamp full of unforeseable squishy obstacles.

This is not the end of surgical craft. The craft changes to one that rewards judgment, oversight, and responsibility over technical virtuosity in suturing or dissection.  While I greatly enjoy the acquisition and display of technical excellence, I know that patients will do better when I can eventually hand off the automated tasks to the robot.