When AI enters hospitals, risks such as "surgical errors and misidentification of organs" also follow.
```
Medical device companies are rapidly integrating AI into surgical and diagnostic equipment, touting “intelligence” as a growth selling point but also introducing new failure modes and liability risks in hospitals, while reports of suspected injuries and malfunctions submitted to regulatory authorities are on the rise.
According to a Reuters review of safety and legal records, and interviews with doctors, nurses, scientists, and regulators, recent reports received by the U.S. Food and Drug Administration (FDA) have included issues such as misleading surgical navigation, undetected arrhythmias, and prenatal ultrasounds “misidentifying body parts.”
In specific cases, healthcare giant Johnson & Johnson’s medical device subsidiary Acclarent announced in 2021 the introduction of machine learning algorithms to its TruDi sinus surgery navigation system.
Reuters, citing unverified FDA reports, stated that after integrating AI, TruDi-related malfunctions and adverse event reports reached at least 100 cases, far exceeding the single-digit pre-AI numbers, with multiple suspected injury lawsuits emerging.
As risks are exposed, regulatory capacity is also under pressure. Reuters quoted five current and former FDA scientists as saying that amidst a surge of AI medical device applications, the FDA found it harder to “keep pace” after losing staff from critical teams.
TruDi Case: Surge in Reports After AI Integration, Lawsuits Point to “Navigation Misguidance”
In 2021, Johnson & Johnson’s Acclarent introduced machine learning to the TruDi Navigation System to help ENT surgeons perform sinus-related surgeries.
Reuters reports that the device added AI functionality about three years after hitting the market. Before AI integration, the FDA had received 7 unverified malfunction reports and 1 patient injury report; after integrating AI, unverified malfunction and adverse event reports reached at least 100 cases.
According to related reports cited by Reuters, at least 10 people were injured between the end of 2021 and November 2025, most incidents allegedly related to TruDi incorrectly indicating the instrument’s position inside the skull to the surgeon. Consequences described in the reports included cerebrospinal fluid leaking from the nasal cavity, accidental perforation of the skull base, and stroke caused by unexpected major artery damage.
Two stroke patients filed lawsuits in Texas, claiming that TruDi system’s AI caused the injuries. One complaint stated that the product might have been “safer before AI integration.” Integra said that “there is no credible evidence” linking the TruDi system, AI technology, and the alleged injuries.
Signals like “Misidentification of Body Parts”: FDA Reports Point to Multiple Types of AI-Enhanced Devices
The FDA emphasizes that reports of adverse events and malfunctions have natural limitations, may lack details, be redacted to protect commercial secrets, or have duplicate reports on the same incident, so attribution cannot rely solely on these reports.
Even so, Reuters’ statistics show that between 2021 and October 2025, at least 1,401 reports submitted to the FDA involved 1,357 FDA-listed AI-powered products (the FDA notes that the list is not exhaustive), with at least 115 reports mentioning software, algorithm or programming issues.
As an example, Reuters noted that a report submitted to the FDA in June 2025 stated that Sonio Detect, used for prenatal ultrasound, had an algorithm problem of “incorrectly marking fetal structures and associating them with the wrong body parts,” although the report did not indicate patient harm. Manufacturer Samsung Medison stated that the report “does not indicate any safety issue.”
Another clue comes from cardiac rhythm monitoring. Reuters notes that at least 16 reports claim that Medtronic’s AI-assisted cardiac monitoring devices failed to detect abnormal rhythms or asystole; these reports did not mention injuries.
Medtronic told Reuters that, after review, the device missed only one abnormal event and “did not result in patient harm,” and stated that some events were related to data display issues rather than the AI itself, while refusing detailed explanations for each case.
Recall Study: AI Device Recall Rate Twice That of All Devices, Defects Exposed Faster
In addition to individual reports, recall data is also strengthening investors’ attention to the “post-market risk curve.”
Reuters, citing a research letter published in JAMA Health Forum in August 2025, reported researchers from Johns Hopkins, Georgetown, and Yale found that 60 FDA-authorized AI medical devices were linked to 182 product recalls, of which 43% occurred less than a year after approval.
The study stated that this recall rate is about twice that of all devices approved under similar FDA regulations.
Approval Path and Safeguards: Most AI Device Approvals Don’t Require Patient Trials, Traditional Framework Questioned
Reuters pointed out that the FDA typically requires clinical trials for new drugs, but medical devices follow a different review path.
Dr. Alexander Everhart, lecturer at Washington University School of Medicine in St. Louis and medical device regulatory expert, told Reuters that most AI-enabled devices entering the market do not need to be tested on patients, but meet regulatory requirements by referencing previously authorized devices without AI capabilities.
Everhart believes that the uncertainty brought by AI is challenging current practices. He told Reuters that the FDA’s traditional regulatory approach to medical devices is “not up to the task” of ensuring AI device safety and effectiveness, and is concerned that in reality, more depends on manufacturers’ self-regulation, raising doubts about whether regulatory safeguards are sufficient.
Regulatory Capacity Under Pressure: Authorized Numbers Doubled, Workload Rises After Team Downsizing
Reuters reports that currently, there are at least 1,357 FDA-authorized medical devices using AI, double the number before 2022.
Reuters cited informed sources as saying that early last year, the Trump administration began dismantling the FDA’s AI team in a cost-cutting campaign led by Elon Musk (DOGE), with about 15 out of 40 AI scientists at DIDSR laid off or choosing to leave, and the Digital Health Center of Excellence, responsible for AI device policy, losing about one-third of its staff, around 30 people.
Some former employees stated that after downsizing, work for some reviewers nearly doubled, and “lack of resources makes it more likely to miss issues.”
HHS spokesperson Andrew Nixon told Reuters the FDA applies the same strict standards to AI-assisted medical devices as other products, stating patient safety is the top priority and that the FDA continues recruiting and training talent in digital health and AI fields.
Risk Disclaimer and Legal NoticeThe market is risky; invest prudently. This article does not constitute personal investment advice and does not consider individual users' specific investment goals, financial situation, or needs. Users should consider whether any opinions, views, or conclusions in this article suit their particular circumstances. Invest accordingly, at your own risk. ```