Florida has launched a criminal investigation into OpenAI and its chatbot ChatGPT following a deadly shooting at Florida State University in 2025.
Attorney General James Uthmeier said the investigation is tied to a review of chat logs between the gunman and ChatGPT.
“This criminal investigation will determine whether OpenAI bears criminal responsibility for ChatGPT’s actions in the shooting at Florida State University last year,” Uthmeier said.
People are just not mentally prepared for the level of sycophancy that these models display.
That kind of large scale gas lighting was previously only available to rich dictators but now OpenAI will rent it to you for $20/mo.
I have a relative that was convinced (by ChatGPT) that his neighbor was beaming radiowaves at him through their window because he had developed pressure sores on the back of his legs (likely from talking to ChatGPT). I brought a Software Defined Radio to his house and left it logging for a week to show him that there was no strong transmissions anywhere in the frequency range that ChatGPT had said.
He told me my equipment was broken. 😑
e: fixed pronouns
Man, that’s sad.
I did, at least, convince him to go to the doctor about the sores (so he would have a paper trail when he sues them).
Smartphones brought way too many people into the worst parts of the Internet without any preparation. Social media’s recommendation system and targeted advertisement made a whole section of our population crazy but LLM chatbots are a whole other level of personalized mental poison.
Right now the ‘evil’ uses of LLMs are just to run online influence campaigns more efficiently. They’re basically taking the old model that required humans to post on social media and providing a force multiplier.
We’re not ready for when these buddy chatbots start pushing political opinions directly. Grok is probably the first example of someone trying that exact thing, but they won’t be the last and they’re only going to get more effective.
