Why Hugo's new AI feature, while seemingly 'unflashy,' is actually patching a critical pain point

What happened?

On July 21, 2026, Medtronic announced a new AI feature for the Hugo surgical robot called the "

Instrument Exit Point," or IEP for short.

When a Hugo instrument moves outside the endoscopic view, the AI detects it and visually indicates where it has disappeared off-screen.

At first glance, it seems like a rather modest feature.

It doesn't automatically recognize blood vessels or teach you about dissection planes.
It doesn't predict bleeding or automatically stop the instruments.

Even so, I don't think we should dismiss this announcement as just another minor convenience added to Hugo.

This is because the problem IEP is trying to address could be a logical patch supported by publicly available FDA data.

The 'invisible risks' of robotic surgery as shown by MAUDE data

To verify the significance of this new feature, I took a slight detour to look at accident data in robotic surgery.

The FDA operates a public database called the "MAUDE (Manufacturer and User Facility Device Experience) database," which collects reports of adverse events related to medical devices.

This is the same massive database I analyzed and introduced in my previous articles when examining da Vinci accidents.

For this article, I directly retrieved data from the FDA's MAUDE API and independently analyzed the latest adverse event reports related to Hugo RAS (2025–2026, 1,251 cases in total) to conduct my own analysis.

The results revealed a very interesting fact.

Hugo-related medical accident information from the MAUDE database

When reading through the details of cases related to "patient injury" (Injury), which was the most serious and frequent type of report, multiple major accidents were found to have been caused by "instrument operation outside the field of view (Out of view / Outside visual field)."

For example, there is the case of a right hemicolectomy reported in June 2026.

The surgeon kept the endoscope pointed at the instruments on the right side while leaving the bipolar fenestrated grasper (BFG) on the left side out of the field of view for approximately five minutes.

Subsequently, moving the BFG without visual confirmation resulted in the forceps perforating the colon.

This caused massive hemorrhaging, leading the patient into shock due to hypotension, and necessitated a transition to emergency laparotomy.

Additionally, in a sigmoidectomy case from May 2026, when the LigaSure RAS (vessel sealing system) was out of view for just under 30 seconds, the moment the assistant moved it to return it to the field of view, it damaged the right iliac artery.

This resulted in two liters of blood loss, requiring blood transfusion and laparotomy.

These cases demonstrate the fact that in robotic surgery, the situation where an instrument is out of view is not merely an inconvenience, but a serious risk that can lead directly to complications.

The package insert for Hugo electrosurgical instruments published by Japan's PMDA also explicitly states that to avoid patient injury, electrosurgical energy should only be activated while under "direct visualization."

This is not a unique problem specific to Hugo, but rather a fundamental principle common to all robotic surgery.

IEP is an attempt to provide the surgeon with auxiliary information to uphold that principle of "direct visualization."

It does not claim to completely prevent adverse events, but is a mechanism to help the surgeon remain conscious of "where the forceps are right now."

IEP is not about "making things visible," but about "making one aware."

Here, one must read the FDA documents carefully rather than letting one's imagination run wild based solely on the term "real-time AI" proposed by Medtronic.

What IEP monitors are the specific energy forceps used with Hugo.

The items listed in the FDA documentation are the following three types: Bipolar Maryland Forceps, Bipolar Fenestrated Grasper, and Monopolar Curved Scissors.

The video is analyzed at approximately 10 frames per second, and when the forceps go out of view, a mark is displayed at the last confirmed position.

However, it is not tracking the current position information of the forceps while they are off-screen.

For example, if the forceps disappear from the right edge of the screen and the surgeon moves them further to the right, the IEP will still only indicate the "location where they last disappeared from the screen."

It is not a feature that estimates the location of the forceps tip within the abdominal cavity or detects contact with tissue.

In other words, IEP is not an 'AI that makes off-screen forceps visible'.

It is an AI that reminds the surgeon of the possibility that they have lost track of the forceps.

This distinction is extremely important.

A surgeon's cognitive load does not arise solely from identifying difficult dissection planes. They are constantly updating multiple pieces of information in their mind, such as camera position, remaining instruments, assistant maneuvers, energy settings, bleeding points, and surgical steps. IEP is an attempt to offload one of those tasks to the machine.

Externalizing cognitive load, similar to 'CAS' in the aviation industry

Let's shift our perspective for a moment.

This 'alerting' approach is very similar to the evolution of

CAS (Collision Avoidance System)

in the aviation industry.

In the past, pilots had to constantly keep track of surrounding aircraft using instruments and visual observation.

However, it is known that in situations with too much information (high cognitive load), humans tend to focus too much on the 'task at hand ('Attention Tunneling')' and overlook dangers outside their field of view.

CAS is a system that monitors the surroundings on behalf of the pilot and alerts them that 'there might be an aircraft over there'.

It does not take control away from the pilot, but rather compensates for cognitive blind spots.

IEP is the same.

In the state of extreme concentration that is surgery, when a surgeon focuses too much on the task at hand, such as 'dissection' (Attention Tunneling), it 'alerts' them to the presence of an energy forceps left off-screen.

This can be called the first step toward the 'externalization of cognitive load' in robotic surgery.

'FDA approved' does not mean 'safety has been proven'

IEP received FDA 510(k) clearance on June 4, 2026 (Submission Number: K253984).

It is classified as a Class II 'software for real-time image enhancement'.

It is important not to misunderstand that FDA clearance does not mean that IEP reduces organ damage or complications.

According to FDA documents, the evaluation included quantitative tests using representative surgical footage, usability tests in simulated operating rooms, and human factors evaluations.

It is stated that the system met pre-set criteria for temporal accuracy in detecting instrument exit and re-entry, sensitivity, specificity, and spatial accuracy in indicating exit positions.

However, the published 510(k) summary does not contain specific numerical values for sensitivity or specificity. Furthermore, no studies comparing clinical outcomes such as complications, accidental electrical contact, or surgical duration in actual patients were presented.

At this stage, all that can be said is that 'under certain conditions, the system detected when the corresponding instrument went off-screen and displayed it in a way the surgeon could understand'.

It cannot yet be said that 'it has been proven to improve patient safety'.

What is truly important is the 'foundation', not the IEP

Looking at IEP alone, its capabilities are limited.

Nevertheless, the strategic significance of this announcement lies in the fact that Medtronic has placed an independent AI platform with sufficient computing power alongside its surgical robot.

Touch Surgery Aide is a computing platform that uses NVIDIA's Holoscan, CUDA, and TensorRT to process surgical video and intraoperative data in real time.

It is designed to run multiple AI applications simultaneously, and Medtronic's Touch Surgery ecosystem is already in use in over 1,500 operating rooms worldwide.

Think about smartphones.

You cannot understand the potential of a smartphone just by looking at the calculator app that came pre-installed.

What was important was the existence of a platform that allowed for the later addition of apps for maps, payments, cameras, and healthcare.

Touch Surgery Aide also has the potential to become an 'operating room app platform' for Medtronic.

In the future, this could evolve into automatic recognition of surgical steps, detection of poor visibility due to bleeding or smoke, warnings for improper use of energy devices, and notifications for surgical delays.

These are not currently approved features, but merely a future vision derived from the platform's design.

In 2020, Medtronic acquired Digital Surgery, a company specializing in surgical AI, data analytics, and digital education.

Touch Surgery Aide is not a product that appeared out of nowhere.

The acquisition strategy from six years ago has finally surfaced in the form of real-time intraoperative AI.

Summary: What the 'unassuming features' are asking us

IEP is the first attempt to use AI to respond to the critical pain point of 'risks of off-screen energy instruments' identified by adverse event data in robotic surgery.

What it can do is limited.
Improvements in clinical outcomes have not yet been proven.

However, the concept of 'the machine taking on some of the surgeon's cognitive load' points to the future direction of robotic surgery AI.

As MAUDE data shows, accidents in robotic surgery do not happen because 'the robot is bad'.

They occur where human cognitive limits and the machine's field-of-view limits overlap.

IEP is trying to make that overlap just a little bit smaller.

It may look unassuming, but it is actually hitting the critical point.

Features that truly add value in the clinical setting are exactly those kinds of features.

Author Profile | Surgical Field Insider

Gastrointestinal surgeon / I specialize in robot-assisted surgery and have made it my life's work to bridge the gap between Japanese medical practice and technological innovation.
I will be writing about the 'realities' that cannot be discussed under my real name here on note.