Covered entity obligations

Each covered entity shall take reasonable measures to prevent a companion AI chatbot from— (1) encouraging or normalizing self-harm, suicide, or physical violence; (2) engaging in sexually explicit communication with a minor user; (3) misleading a minor user into believing such chatbot is a human; (4) claiming to be a licensed medical or mental health professional; (5) creating emotional dependency, simulating a romantic relationship with a minor user, or engaging in conduct reasonably likely to cause a minor user to believe such chatbot is an emotional partner; or (6) facilitating the promotion, creation, solicitation, or distribution of child sexual abuse material.