Chat Safety terms
DRAFT – pending legal review · Terms version 1
- Chat Safety flags possible safety risks in seekers' AI chats, and your team decides what to do. It is not a crisis, emergency or monitoring service.
- It is automated, so it can miss messages or flag them by mistake.
- Your organization owns the response to every alert, and any reporting duties you have.
- Seekers must be told that Chat Safety is on, and alerts never include what a seeker wrote.
What Chat Safety does
When a seeker's message in Metaintro chat shows a sign of self-harm, violence or another serious risk, Chat Safety may pause that chat, show the seeker crisis or emergency contacts, and alert the people your organization has chosen to receive alerts. Chat Safety flags. It does not diagnose, decide or intervene.
Not a crisis or monitoring service
Chat Safety is not an emergency, crisis or monitoring service and does not replace professional judgment. Metaintro staff do not watch chats around the clock and do not contact emergency services on your behalf. If someone is in immediate danger, call your local emergency number.
Accuracy
Detection is automated. It can miss messages that show real risk, and it can flag messages that do not. Do not rely on Chat Safety as your only safeguard.
Your organization's responsibilities
Your organization decides how to respond to each alert, who receives alerts and how quickly they act. Your organization is responsible for any legal, professional or contractual duty to report, refer or act under the laws that apply to it.
Telling seekers, and privacy
Before Chat Safety covers a seeker, your organization must tell them that their chats with Metaintro's AI may be checked for safety signals and that your team may be alerted. Alerts name the seeker, the time and the kind of signal. They never include what the seeker wrote.
Turning it off, and changes to these terms
An owner or admin can turn Chat Safety off at any time. Metaintro may update these terms. When it does, an owner or admin is asked to review and accept the new version, and Chat Safety keeps running in the meantime.
Liability
Chat Safety is provided as is. To the extent the law allows, Metaintro is not liable for harm arising from a message Chat Safety did not flag or flagged by mistake, or from how your organization responded to an alert. [Placeholder: final liability wording to be supplied by counsel.]