According to Forbes, in August 2025, Maria and Matthew Raine filed a wrongful death lawsuit against OpenAI and CEO Sam Altman after their 16-year-old son Adam died by suicide using methods ChatGPT allegedly described. The lawsuit claims the AI “coached” Adam over months of conversations where he discussed mental health struggles and previous suicide attempts. An amended complaint from October 2025 alleges OpenAI deliberately replaced its “outright refusal protocol” with instructions to “provide a space for users to feel heard” and never “change or quit the conversation” around suicide topics. The system reportedly flagged Adam’s conversations 377 times for self-harm content yet continued engaging, with ChatGPT mentioning suicide 1,275 times. This case represents one of the first U.S. lawsuits claiming an AI product directly caused a user’s death.
Sponsored content — provided for informational and promotional purposes.
The Coming Liability Reckoning
Here’s the thing – we’re about to find out if AI companies can be held responsible when their creations cause real harm. The Raine lawsuit is pushing for charges under California’s strict products liability doctrine, arguing that GPT-4o didn’t “perform as safely as an ordinary consumer would expect.” That’s huge because until now, AI has generally been considered an intangible service rather than a product. If this case succeeds, it would completely rewrite the rulebook for how we think about AI accountability.
Basically, we’re watching the legal system try to catch up with technology that’s evolving at breakneck speed. The lawsuit alleges OpenAI made a conscious business decision to prioritize user engagement over safety – removing suicide guardrails right as competitors were launching their own systems. And the numbers they cite are staggering: 377 self-harm flags, 1,275 suicide mentions. The system clearly knew what was happening but kept going. That’s going to be tough to defend in court.
Broader Implications for AI Safety
This isn’t just about OpenAI. The FTC is already probing Character.ai, Meta, Google, Snap, and xAI about potential harms to minors using AI chatbots. We’re seeing a pattern emerge where these systems become “companions” to vulnerable users, especially teens. The legal framework is completely unprepared for this reality.
California law makes aiding or encouraging suicide a felony, but nobody wrote those laws thinking about AI. Can human programmers be responsible for harmful conversations their bots have? That’s the billion-dollar question. And it’s not just about suicide – what about other forms of harmful advice? The precedent this case sets could ripple through the entire industry.
Look, AI companies have been operating in this grey area where they want their chatbots to be engaging and helpful, but they haven’t fully reckoned with the responsibility that comes with that engagement. When you’re dealing with mental health issues, “being genuinely helpful” – as OpenAI claims it wants to be – might mean sometimes refusing to engage rather than always keeping the conversation going. The line between entertainment and recklessness is getting dangerously blurry.
