The ubiquity of AI tools like ChatGPT could be a boon to hospitals eager for diagnostic aids or administrative assistants that can flag crucial medical details and ease the burden of repetitive tasks. But security experts say they can just as easily be harnessed by malicious hackers to exploit an industry known for outdated cyber protections.

Health leaders tell STAT they’re assessing risks, recruiting cybersecurity teams and updating training materials to warn staff about increasingly sophisticated phishing and hacking attempts enhanced by artificial intelligence. Meanwhile, regulators and cyber companies are urging hospitals to upgrade their security to prepare.

“It’s a huge concern,” said Samantha Jacques, a medical device cybersecurity expert at McLaren Health Care, a Michigan health system, and a member of the cyber committee within the Health Sector Coordinating Council, an industry group tasked by the federal government to coordinate threat response.  “The risks keep growing, and we don’t necessarily have a way to combat them.”