Medical safety concerns highlighted: Limitations exist in the application of artificial intelligence

With the development of artificial intelligence technology, medical device manufacturers are scrambling to integrate AI into their products. However, as more medical devices incorporate AI, patient injury claims caused by AI continue to rise. According to existing reports, after Johnson & Johnson's Acclarent incorporated AI into its TruDi sinus surgery navigation system, adverse event reports surged from 8 to at least 100. Between the end of 2021 and November 2025, at least 10 people were injured, including serious cases such as strokes. Related lawsuits allege that AI reduced the product's safety, while the acquirer denies any causal relationship.

Proponents of AI in healthcare aim to use technology to improve diagnostic and treatment efficiency, solve medical challenges, and empower both doctors and patients. Meanwhile, manufacturers are rushing to deploy AI medical devices primarily to seize market opportunities, pursue technological dividends, and generate profits, with some companies even lowering safety standards to speed up market entry. The FDA is supposed to fulfill its regulatory duties and ensure patient safety, but under the Trump administration’s cost-cutting measures, its core AI regulatory team was dismantled, and insufficient manpower has made it difficult to handle the massive demand for device approvals and monitoring. The companies involved deny any link between AI and harm, essentially to protect their brand reputation and avoid legal and financial liability.

The safety risks of AI medical devices have caused multiple negative impacts. Patients have suffered serious injuries due to AI device malfunctions, their health rights have been compromised, and public trust in AI medical devices has been shaken. Secondly, AI devices have a high recall rate, and adverse event reports suffer from issues such as missing data and inadequate classification, which not only disrupt industry order but also hinder the healthy development of AI medical technology. In addition, the FDA's regulatory capacity is insufficient, and traditional regulatory models are outdated, which not only undermines its credibility but also exposes the shortcomings of the existing regulatory system in adapting to AI technology iterations, making it difficult to establish effective safety safeguards. At the same time, the improper use of generative AI chatbots poses new risks and challenges to medical diagnosis and treatment.

To mitigate the impact of medical AI, coordinated efforts from multiple parties are required. Manufacturers should strengthen their primary responsibilities, strictly control safety standards, improve AI algorithm testing, and avoid lowering safety thresholds for the sake of market competition, while cooperating with regulatory authorities to monitor post-market performance and report adverse events. Secondly, the FDA should optimize the regulatory system, supplement with AI regulatory experts, improve the adverse event reporting system, establish a full lifecycle regulatory framework tailored to AI characteristics, and refine approval standards to avoid relying on traditional device approval logic. At the industry level, standards should be established to enhance self-regulation, promote the standardization of AI medical device training data, and reduce algorithmic bias and instability. Finally, expert oversight and public science education should be strengthened, utilizing professional expertise to improve technical standards and guide patients and healthcare providers to view AI in medicine rationally, thereby avoiding risks from improper use.

In summary, AI medical technology should ideally serve as a catalyst for the advancement of medicine. However, the profit-driven strategies of manufacturers and the lagging regulatory system have led to prominent safety risks, resulting in a dilemma of "imbalance between innovation and safety." The FDA's insufficient regulatory capacity, the inability of traditional regulatory models to adapt to AI characteristics, and some companies' disregard for safety in their rush to achieve results have collectively exacerbated industry chaos. In the future, only by adhering to the principle of "safety first, orderly innovation," strengthening manufacturers' primary responsibilities, promoting the FDA to improve the regulatory system and enhance its capabilities, and uniting the efforts of the industry, experts, and the public, can safety risks be mitigated. This will enable AI medical technology to truly empower medical development, unlocking technological potential while effectively ensuring patients' lives and health.