Modern AI tools are helping people across the world, but they aren’t built to recognize when a user is in emotional distress. Without a way to detect and respond to serious mental health red flags, users who need real human support may be left unheard or worse, unintentionally harmed.

What We’re Asking:

We believe AI developers should build in a gentle, private system that allows AI to say:
“Would you like to talk to a human?”

This doesn’t mean surveillance or publicly reporting users chats. It means compassion.

We’re asking leading AI companies—OpenAI, Google DeepMind, Anthropic, xAI, Mistral, Cohere, and others to:

Include a baseline model of emotional wellbeing so the AI can recognize concerning language shifts.

Gently offer a human help option to users who may be in distress, without alerting authorities.

Provide in-house mental health support or partner with licensed services, staffed by professionals under confidentiality law.

Why This Matters:

Millions of people use AI when they have no one else to talk to.
We must create systems that don’t just mirror their distress—but respond with care.
This change could help prevent harm, save lives, and build public trust in a new era of human-AI collaboration.

Sign now to ask for AI that helps, not harms.