CHAI forms health AI cybersecurity work group to draft frontier risk guidance
Coalition for Health AI said it is forming a new work group to develop cybersecurity playbooks and tools for healthcare organizations by the end of 2026.
Coalition for Health AI said on Aug. 12 that it is forming a Health AI Cybersecurity Work Group to develop playbooks and related guidance on frontier model cyber risk and readiness for healthcare organizations.
The nonprofit said the effort is aimed at health systems, payers and health tech companies. CHAI said the work is meant to help those groups respond to frontier model threats and use frontier models for defense.
CHAI said the work group will be led by a leadership council of health system, payer and industry experts and will include nearly 100 members. The group is set to meet in bi-weekly sessions and produce a defensive playbook, an offensive playbook, a frontier AI cyber risk assessment tool and other resources.
The organization said it expects the playbooks at the end of 2026 and described the effort as the sector’s first practical, peer-developed guidance on frontier model cyber risk and readiness.
In its release, CHAI cited Anthropic’s Mythos-class frontier models and said the public Fable variant was released on June 9, 2026. The release said those capabilities changed the cybersecurity threat and defense landscape for healthcare by compressing attack timelines, industrializing protected health information exfiltration and ransomware, and improving vulnerability management and incident response for defenders.
CHAI CEO Brian Anderson said the organization is convening cyber experts from community hospitals to academic medical centers and the technology and cybersecurity ecosystem to keep patients, staff, data and operations safe.
John Flores, CISO of the University of Texas Medical Branch, said AI advances change the health care cybersecurity threat level and that hospitals need a real toolkit.
Isaiah Nathaniel, SVP and CISO of Delaware Valley Community Health, said the new speed, scale and sophistication of cyberthreats require a new playbook.
CHAI said the work group builds on its broader mission of developing consensus-driven frameworks, governance tools and best practices for trustworthy AI adoption in healthcare. The group’s next milestones are its bi-weekly sessions and the planned release of its playbooks by the end of 2026.
Sources
Editorial note: This article was prepared with AI assistance and includes source attribution for reader review.
View the Web StoryA visual, tap-through version of this reportThis is the AMP version of the article. Read the full article on Atlas Watchline.