Epic Systems identified software flaws in certain MyChart configurations that could allow unauthorized individuals to view patient records without appearing in audit logs, using Anthropic's Claude Mythos model.

Overview Debates Related claims Topics Sources

Epic Systems chief security officer Stirling Martin participated in Anthropic's Project Glasswing program to test how attackers could utilize open-source artificial intelligence agents to access confidential health records.

Bank of England Governor Andrew Bailey described frontier AI threat capabilities as the most immediate cyber risk facing global financial systems in an August 31, 2026 letter.

Dr. Sean Kelly of Imprivata stated that an autonomous AI agent with excessive permissions could access sensitive patient information, enter incorrect medical record data, or alter medication dosages.

Dr. Sean Kelly of Imprivata warned that because AI agents operate autonomously at machine speed, a single error could propagate before being recognized, directly impacting care delivery.

Dr. Sean Kelly of Imprivata recommended that AI agents have clearly defined identities, permissions, and boundaries, with human review for higher-risk activities and audit trails.

The Imprivata survey showed that more than 85% of healthcare leaders responsible for AI strategy are confident they have visibility into AI agent activity and can fully control autonomous actions.

The Imprivata survey found that more than half of healthcare leaders ranked security among their top concerns when adopting agentic AI.

Security experts state that agentic AI presents security risks because agents can act autonomously, traverse different information systems, and execute actions on behalf of users.