When do we need to tell users they are interacting with AI?
Use different triggers for the EU and Utah. A single generic AI notice misses the distinction.
Tell users at first interaction
Covered direct-interaction systems need a clear AI disclosure by the first interaction, unless the context makes this obvious.
Separate two triggers
Covered consumer transactions require disclosure when clearly asked. High-risk regulated interactions require prominent upfront disclosure.
Who provides the AI system, and who deploys it?