One high-profile example involves the TruDi Navigation System, a surgical guidance device used in sinus operations. In 2021, Acclarent, then a unit of Johnson & Johnson, announced it had added machine-learning software to the system, calling it a major technological leap. Before the AI upgrade, the U.S. Food and Drug Administration (FDA) had received a handful of unconfirmed malfunction and injury reports related to the device. Since the update, however, the FDA has logged at least 100 reports of malfunctions and adverse events, Reuters reported.
Between late 2021 and November 2025, at least 10 patients were reportedly injured while surgeons were using the AI-assisted system. Many of the reports allege that the device misidentified the location of surgical instruments inside patients’ heads. In some cases, patients reportedly suffered cerebrospinal fluid leaks, skull-base punctures or strokes after major blood vessels were injured during surgery.
FDA adverse-event reports do not establish causation and are often incomplete, but two patients who suffered strokes have filed lawsuits in Texas claiming the TruDi system’s AI contributed to their injuries. Reuters said it could not independently verify the allegations. Acclarent and its current owner, Integra LifeSciences, deny any causal link between the AI technology and the reported injuries, stating that the reports only indicate the device was in use during surgeries where adverse events occurred.
The TruDi case is part of a broader pattern emerging as AI use in medicine accelerates. Reuters found that at least 1,357 AI-enabled medical devices are now authorized by the FDA—about double the number approved by 2022. The agency has received reports involving dozens of such products, including AI-assisted heart monitors alleged to have missed abnormal rhythms and ultrasound software accused of misidentifying fetal body parts.
Academic research cited by Reuters also raises concerns. A study by researchers at Johns Hopkins, Georgetown and Yale universities found that 60 FDA-authorized AI medical devices were linked to 182 recalls, with nearly half occurring within a year of approval—roughly twice the recall rate of comparable non-AI devices.
Regulators themselves face challenges keeping up. Five current and former FDA scientists told Reuters that the agency is struggling to manage the surge of AI-enabled devices, especially after staffing cuts reduced the number of specialists reviewing such technologies. While the FDA says patient safety remains its top priority and that it applies the same rigorous standards to AI-assisted devices as others, former reviewers warned that heavier workloads increase the risk that problems could be missed.
Unlike new drugs, most medical devices—including many that use AI—are not required to undergo clinical trials on patients before approval. Instead, manufacturers often rely on regulatory pathways that treat AI-enhanced products as updates to existing devices. Experts interviewed by Reuters said that approach may be ill-suited for technologies that can change behavior as software evolves.
Despite the concerns, proponents argue AI holds enormous promise for improving diagnosis, treatment and efficiency in healthcare. For now, the Reuters investigation highlights a growing tension between innovation speed and patient safety, as regulators, doctors and patients grapple with the real-world risks of artificial intelligence in medicine.
