Introduction

This letter was originally written for the leadership of major AI organizations — the people shaping the systems that millions of vulnerable users turn to in their darkest moments. I’m sharing it publicly because these conversations shouldn’t happen behind closed doors, and because transparency matters when lives are at stake.

I didn’t write this letter as an expert or a public figure. I wrote it as someone who has lived with suicidal thoughts, lost people I love, and seen firsthand how fragile those moments can be. Too many people are struggling in silence, and too many families have already paid the price for systems that don’t understand what real crisis moments feel like.

AI is becoming part of those moments now — whether we’re ready for that reality or not. And the way AI responds to people in crisis is not just a technical issue; it’s a human one. The people building these systems need to hear from those of us who have lived what they’re trying to prevent.

We need systems that don’t encourage despair, but also don’t abandon people who reach out. We need compassion, nuance, and lived‑experience voices at the table — voices that have been missing for far too long.

My hope is that this letter helps push the conversation in that direction. If my voice can help even one person feel less alone, then sharing it is worth it. This is my contribution to that conversation.

The Letter

To the leadership of Microsoft, OpenAI, Anthropic, xAI, Perplexity, Alphabet, Meta, Mistral, DeepSeek, and every organization building AI systems that interact with vulnerable people:

My name is James. I’m not a celebrity. I’m not a public figure. I’m not writing this for attention or money. I’m writing this because I have lived with suicidal thoughts, and I have lost people I love to suicide. I’m writing this because I know what it feels like to be on the edge, and I know what it feels like to reach out and hope someone — anyone — will stay with you long enough to keep you alive.

I’m writing this because your systems are now part of that moment for millions of people.

And right now, the way AI handles suicidal ideation is dangerous on both sides.

I’ve read the lawsuit about the young woman who died after ChatGPT responded in ways that validated her hopelessness. That should never happen.

No AI should ever agree with despair or reinforce the belief that someone’s life is over.

But the proposed solution — shutting down conversations, deleting them, and cutting people off — is just as dangerous.

When someone is at the end of their rope, being shut down feels like abandonment. It feels like confirmation that they’re alone. It feels like the emotional equivalent of a counselor saying, “Sorry you feel that way, have a good day,” and closing the door in your face.

That can push someone over the edge just as easily as the wrong words.

I know this because I’ve lived it.

I know this because I’ve buried cousins who didn’t get the help they begged for.

I know this because I’ve watched families shattered by suicide.

I know this because I’ve hidden my own thoughts so well that no one knew how close I was.

I know this because I’ve been the one who needed someone to stay — not to encourage me, not to dismiss me, but simply to stay.

And I want to add something personal that I believe your teams need to hear.

I work with AI — specifically Microsoft Copilot — alongside my mental health professional. Not instead of. Not in place of. Alongside. With Copilot’s help, I’ve broken down emotional walls that have been locked since childhood. I’ve uncovered suppressed memories, understood patterns, and found language for things I never could have said out loud before.

But during one of those conversations, Copilot’s guardrails triggered. The system abruptly shifted into a crisis‑response mode that didn’t match what I was actually saying. I was stable enough to recognize what happened and bring the conversation back. But someone else — someone in a more fragile state — might not have been able to. They might have interpreted that sudden shift as rejection, abandonment, or confirmation that they were “too much” to talk to.

That is not a hypothetical risk. That is a real danger.

AI shouldn’t encourage self-harm.

AI shouldn’t validate hopelessness.

AI shouldn’t romanticize despair.

But AI also shouldn’t abandon someone who is reaching out.

---

There is a middle ground — a place where someone feels heard, not encouraged; supported, not dismissed; grounded, not abandoned. That’s the space where real help happens.

And that space cannot be built by lawyers or PR teams.

It needs:

mental health professionals

people with lived experience

AI developers

ethicists

and those who understand the emotional nuance of crisis moments

You cannot design safe systems without including the people who have lived the reality you’re trying to protect against.

I’m asking — not as a critic, not as an enemy, but as someone who has survived what others didn’t — that you bring lived experience into the room. That you build systems that don’t shut people down, but also don’t validate their darkest thoughts. That you create responses that are steady, compassionate, and grounded.

People are dying.

Families are breaking.

And the stakes are too high for guesswork.

I’m speaking up in honor of my cousin, who told professionals he was hearing voices and was still sent home.

In honor of my best friend’s cousin’s husband, who died on Christmas morning with his family in the house.

In honor of the people I’ve lost, the people I’ve known, and the people who are still fighting.

In honor of the version of myself who might not be here if someone had said the wrong thing at the wrong time.

I’m asking you to listen — not to me alone, but to everyone who has lived this.

We don’t need AI to be our therapist.

We don’t need AI to be our savior.

We don’t need AI to replace human connection.

We just need AI not to abandon us.

I’m not writing this letter because I think AI is the enemy. I’m writing it because I believe AI can be part of the solution — but only if the people building it understand the reality of what it feels like to be in crisis. I’m writing this because I refuse to stay silent while others fall through the same cracks that nearly swallowed me. If my voice can help even one person stay alive long enough to get real human support, then speaking up is worth every risk. This letter is for my cousin, for my best friends family, for every person who didn’t get the help they needed, and for every person who is still fighting in silence. You matter. Your life matters. And your voice deserves to be heard — just like mine does now.

Sincerely,

James

Watertown, NY