Oregon AI chatbot law

The enrolled measure is titled as relating to artificial intelligence companions. The legislative digest says operators must notify users that they are interacting with artificial output when a reasonable person would believe they were interacting with a natural person. It requires a protocol for detecting suicidal ideation or intent, or self-harm ideation or intent, and for preventing output that could cause that ideation. The digest says the protocol's minimum contents include referral to an appropriate crisis lifeline and further steps informed by clinical practice.

If the operator has reason to believe the user is a minor, the measure requires additional statements, including steps so the companion does not lead a reasonable person to believe they are talking with a natural person. Operators must post a yearly public report of how often they referred a user to resources to prevent suicidal ideation, suicide, or self-harm. A harmed user may sue. The governor approved the measure on March 31, 2026. Oregon Laws 2026, chapter 85, states an effective date of January 1, 2027. Washington uses that same calendar day and is still a different statute.

Oregon is explicit about suicide and self-harm detection and about a public count of referrals. That is a safety design, and it is also a clinical claim a product should not fake. A protocol informed by clinical practice is not the same sentence as a chatbot that says it is a therapist.

These state laws are not copies of each other. A duty in Georgia may be missing in Tennessee, and a date in one state is not a date in the next. Whether Congress will override some of them is a live debate, described on the federal AI preemption page. Until a federal statute actually does that, read the state text that applies to you.

Oregon's chatbot law requires a notice when a chat feels human and a protocol for suicidal ideation, effective January 1, 2027. HeyOtto is not a clinician and does not claim Oregon's detection protocol. If a child shows distress, the chat shows crisis resources, including 988, and the parent gets an alert.

• Parents can read chats in the parent dashboard.
• Distress produces an alert and crisis resources, not an automatic call to emergency services.
• HeyOtto does not claim this statute is already satisfied.