The Coalition for Health AI (CHAI) announced the formation of a working group comprising approximately 100 leaders from healthcare systems, insurance companies, and the industrial sector to address cybersecurity risks associated with advanced AI models, known as Frontier AI. The group will meet every two weeks with the goal of developing and publishing cybersecurity guidance for healthcare AI before the end of 2026.
The group’s outputs will include an AI risk assessment tool, along with guidance documents addressing defensive and offensive security strategies. Its formation comes at a time when CHAI believes that recent developments in advanced models have fundamentally changed the nature of the cyber threats facing the healthcare sector.
More Capable and Autonomous Models
The alliance said that Anthropic’s launch of the advanced AI models Mythos and Fable during the spring was a motivating factor behind the creation of the working group. Fable is the publicly available version, while Mythos, according to the article, can independently detect cybersecurity vulnerabilities and turn them into working exploits.
These models belong to the most powerful AI systems currently available, as they can handle complex tasks, process enormous amounts of information, and operate autonomously without continuous human direction. This gives healthcare organizations an opportunity to improve their defensive capabilities, but it also enables attackers to increase the speed, scale, and effectiveness of attacks.
Participation from Healthcare and Cybersecurity Leaders
John Flores, chief information security officer at University of Texas Medical Branch, is participating on the 14-member leadership council guiding the working group’s efforts. He is joined by representatives from other health systems, including Baptist Health and Duke Health, as well as executives and specialists from healthcare technology and cybersecurity companies, threat intelligence-sharing organizations, and standards-setting bodies.
Flores said that healthcare organizations have always faced cybersecurity challenges, but current developments in AI have changed the level of threat they face. Isaiah Nathaniel, senior vice president and chief information security officer at Delaware Valley Community Health and a member of the leadership council, also stressed the need to equip all parts of the healthcare sector, including smaller systems, to deal with the negative aspects of technological progress.
Growing Pressure on Healthcare Organizations
Healthcare organizations have faced a large number of cyberattacks and data breaches in recent years. These incidents can disrupt patient care, such as by diverting ambulances and canceling appointments and surgeries, or taking electronic health record systems offline.
According to Fortified Health Security, hospitals remediated only 6% of identified cybersecurity risks during the first quarter of 2026, compared with 23% in the first quarter of 2025. These figures indicate that the volume of discovered vulnerabilities exceeds healthcare organizations’ ability to remediate them, increasing the importance of developing guidance and tools to help them assess and respond to AI model risks.