CMS Pledge Sets Voluntary Goals for Patient-Facing AI Assistants

A voluntary pledge from the Centers for Medicare & Medicaid Services (CMS) calls for conversational AI assistants to securely access relevant health information with patient consent, while helping users understand when to contact a health professional.

Anthropic, Google, Microsoft AI and OpenAI are listed as early adopters. The pledge calls for assistants to provide personalized support, distinguish educational information from clinical guidance, and direct patients to health professionals when needed.

On July 30, 2025, CMS launched the wider health technology effort at a White House event. More than 60 companies pledged to collaborate, including 30 that committed to promoting health outcomes through technology. CMS described conversational assistants as potential tools for checking symptoms, navigating care options and scheduling appointments.

The effort followed a May 2025 request for information from CMS and the Assistant Secretary for Technology Policy. Patients, caregivers, providers, payers and technology developers submitted nearly 1,400 comments to CMS in just over a month.

The initiative also aims to improve interoperability, making it easier for information to move between patients and providers, and to make personalized digital tools more available. CMS’ public materials describe voluntary goals and do not establish a binding standard or product requirement for AI assistants.