AI in Hospitals: Between Innovation and Cybersecurity Risks

Artificial intelligence is no longer just a vision of the future in hospitals; it’s happening now. Imaging systems read scans faster than doctors, predictive models flag patients at risk before symptoms appear, and robots assist in surgeries and manage hospital logistics.

But this leap forward comes with a price. Every smart system is also a potential doorway for hackers. Experts warn that cybercriminals could manipulate medical data, inject malware into AI tools, or disrupt systems at critical moments. In some cases, delayed diagnoses or canceled surgeries could mean life or death.

Over the past few years, several incidents highlighted the severity of the threat. In May 2025, the U.S. healthcare network Kettering Health was hit by ransomware, forcing cancellations of medical procedures. In February 2024, Change Healthcare faced one of the largest cyberattacks in the insurance sector by the BlackCat / ALPHV group. In Canada, Covenant Health reported a breach affecting thousands of patients, while the UVM Health Network in Vermont lost over $63 million after an attack disrupted digital infrastructure and patient care.

To reduce these risks, experts propose practical solutions: encrypting all medical data, segmenting networks so a single breach doesn’t shut down entire systems, enforcing multi-factor authentication, and requiring AI vendors to pass strict security audits. Hospitals are also urged to conduct regular penetration tests, use AI to monitor for unusual activity, and train staff to avoid common errors like weak passwords or clicking on phishing links.

The equation is simple: AI can save lives, but it can also create new dangers if left unprotected. The future of smart healthcare must be not only innovative, but also secure.