AI Disclosure

Effective date: August 2, 2026

1. Everything About the Avatar Is Generated

The avatars are software. There is no person on the other end.

• The face is synthetic. Where an avatar is built from a photograph, the moving image is generated from it, not filmed.
• The voice is synthetic, produced by a speech model. Where a voice was cloned from a sample, the audio you hear is generated, not a recording of the person speaking.
• The words come from a language model. Anything that sounds like memory, feeling, opinion, or first-hand experience is generated text. The avatar has none of those things.
• The character is a configuration: a name, a look, a voice, and a written brief. A persona presented as a friend, coach, or companion is still a configuration.

Generated output can be wrong. It is not medical, mental-health, legal, or financial advice, and it is not a substitute for professional care, therapy, or another person. The Terms of Service say the same thing in the place where it binds.

2. How We Disclose It

On Casola’s own surfaces:

• Throughout the session, the name plate on the video reads “AI” next to the avatar’s name, so the label is on screen for as long as the conversation is.
• In our marketing, we do not describe an avatar as a person and we do not present a generated clip as recorded footage.

If an avatar ever tells you it is human, it is wrong, and we would like to know: safety@casola.ai.

3. Why This Is Also the Law

Disclosure is not only good practice; a growing body of law requires it.

• In the EU, transparency rules oblige providers and deployers of AI systems that interact with people to make the artificial nature clear and to mark synthetic audio and video. Those obligations phase in from August 2026.
• A growing number of US states now regulate companion chatbots directly — California and New York among the first — requiring a clear disclosure that the user is talking to AI, reminders during longer sessions, and a self-harm and crisis protocol.
• US consumer-protection law treats giving users the deceptive impression that they are talking to a human as an unfair or deceptive practice, independent of any AI-specific statute.

4. If You Build on the API

The disclosure duty in your product is yours, not ours. You control the interface, so you are the deployer, and Section 5 of the Developer & API Terms makes it a condition of use.

What we would do in your position:

• Say it before the microphone or camera turns on, in the interface, not in a linked policy.
• Say what is synthetic: face, voice, and words. “Powered by AI” does not tell anyone that the voice they are about to hear is generated.
• Repeat it in long sessions, and whenever the conversation resumes after a break.
• Instruct the model never to claim to be human and never to claim credentials, and test that it holds under pressure.
• Never present a generated likeness of a real person without saying it is generated.
• If you record, say so separately. A recording notice is not an AI notice and neither one covers the other.

This is guidance from an infrastructure provider, not legal advice. Your obligations depend on where your users are and what your product does, and you should get advice on both.

5. Contact

Questions about this page: legal@casola.ai. Safety concerns: safety@casola.ai.