AI in healthcare faces risks from deepfakes and synthetic data

The rapid integration of artificial intelligence in healthcare is presenting both clinical opportunities and significant risks regarding data integrity. While AI supports diagnostics, clinical documentation, and patient chatbots, the rise of “synthetic patients” and “deepfake” medical imagery poses a threat to clinical research and medical guidelines.

According to WHO/Europe, nearly two-thirds of countries in the region use AI in diagnostics, yet only 8% have a dedicated national strategy for healthcare AI. In the United States, the FDA maintains a list of AI-enabled medical devices across specialties like radiology and cardiology. In Italy, a significant portion of specialists and general practitioners are already utilizing these technologies.

A critical concern is the emergence of fraudulent medical data. Generative models can create entirely fabricated clinical identities, including names, lab values, and consistent medical histories. Research has shown that even trained radiologists struggle to detect these forgeries; in one study, spontaneous recognition of fake X-rays was only 41%. This capability to produce high-quality, fake medical records threatens the validity of clinical trials, insurance processes, and evidence-based medicine.