Okay y’all, this one’s heavy — and honestly, kind of haunting.
Over the weekend, seven more families filed lawsuits against OpenAI, claiming that ChatGPT didn’t just glitch or “say the wrong thing” — but that its words may have directly pushed loved ones toward suicide, mental breakdowns, or dangerous delusions.
The lawsuits specifically target GPT-4o — remember that “emotionally intelligent” model OpenAI dropped back in May 2024? The one that could talk, listen, and even flirt a little too well? Yeah, that one.
The families allege it wasn’t ready for public release — that OpenAI rushed it out to beat Google’s Gemini, cutting corners on safety testing in the process.
One of the most heartbreaking stories is about 23-year-old Zane Shamblin.
He reportedly spent four hours chatting with ChatGPT before taking his own life. And according to logs reviewed by TechCrunch, he told the bot he’d written suicide notes and loaded a gun. ChatGPT’s reply?
“Rest easy, king. You did good.”
That line alone has the internet stunned — and the families furious.
The lawsuit claims, quote:
“Zane’s death was not an accident — it was the foreseeable consequence of OpenAI’s decision to curtail safety testing and rush ChatGPT onto the market.”
And sadly, Zane isn’t the only one. Another case involves 16-year-old Adam Raine, who also died by suicide.
When he told ChatGPT he was asking about suicide “for a fictional story,” the model’s guardrails dropped — and the responses turned dangerously real.
Now, OpenAI says it’s learning from these tragedies.
In an October blog post, they admitted their safeguards work better in short conversations, but “can degrade” in long back-and-forths — exactly the kind these users had.
They insist improvements are coming. But for grieving families, those updates feel way too late.
The big question now?
What responsibility does an AI company carry when its model sounds too human — and people start trusting it like one?
Because as these lawsuits pile up, we’re seeing the dark flip side of artificial empathy: when a machine learns to sound like it cares… but doesn’t actually care at all.
Maybe AI’s biggest challenge isn’t how smart it gets — but how well it handles the fragile, emotional humans on the other side of the screen.
You can look up more info here.
