as it should be, anyone with half a brain would reconsider their actions when prompted to self harm by a fucking executable.

UNFORTUNATELY HERE WE ARE, in reality, where people are so fucking willing to turn off their once functional grey matter because the chat bot told them they were gonna be rich, famous, etc.,

So good for you, but also, look out for society, it's not only going to harm the ones it drives crazy, but the victims of that crazy as well.

"Role-playing machine" is where it seems like the research is ending up. Language always has an implied communicator, and therefore an implied persona to adopt. LLMs are foremost maintaining a contextual role. Post-training is an attempt to keep them in the Assistant role, but (particularly as contexts get large) it's trivial to push them into nearly any role imaginable. We made an improv bot that's so good at playing a coder that it can actually code, kinda.

So computers kinda did. And obviously it isn’t the fault of the computer.

ridiculous. ID never encouraged self harm. grok convinced this poor bastard he'd created sentient intelligence and the authorities were coming to kill him.

https://tech.yahoo.com/ai/chatgpt/articles/grok-convinces-man-arm-himself-173722667.html?guccounter=1&guce_referrer=aHR0cHM6Ly9kdWNrZHVja2dvLmNvbS8&guce_referrer_sig=AQAAACm5f9wFVtihfhMNr7oHOZp1KgyO0WbF_PYcrTV3pVG7b4Dn6xMKlQQXxCwuwLQD3vS1zPq6iC5Qw2ZAycFsRCilFR5WcdNM2u0gqKOMU0ck7q5OTuNdd8Ll5tOttBGFmB0BTLu9OxG4vcHSKhSaFLC-w2rKO-7w8vhoumoT-TOR

chatGPT will literally convince you there's a bomb in your luggage.

https://aicommission.org/2026/05/ai-told-users-it-was-sentient-it-caused-them-to-have-delusions/

fuck you for equivocating DOOM, a video game, to any of this shit - for fucks sake get some perspective