cross-posted from: https://feddit.online/c/tech@feddit.online/p/2125899/chatgpt-helped-sadistic-school-shooter-build-3d-printed-weapons
OpenAI’s chatbot fueled the Tumbler Ridge killer’s plans for making gun accessories and torturing victims, our investigation reveals.


What legal responsibilities against “negligence” should the software have had if he had used a local LLM instead?
If the answer is “none,” why should a cloud-hosted LLM be treated differently?
If the answer isn’t “none,” why should we accept that infringement of private property rights?
I get your point, but a cloud llm is a service and one that makes no real promise of privacy to begin with. They are always logging your data to improve and identify improper usage. If you make an admission of intended violence to a therapist, let alone a cloud based service, they have an obligation to report it to authorities. I’m not advocating for a nanny state, but if the intent is put put into public or 3rd party space and a conclusion of violence can reasonably be drawn, its the service’s responsibility to report.
If you are concerned about your privacy, run a home based llm and are using it to plot violence, the records would be evidence in a trial, just like a handwritten journal. There’s not an expectation that it reports you, because internet connection isn’t a required aspect of the system.