Then again, people willing to make these comments aren't people that I tend to understand or empathize with anyway, so using a chatbot shouldn't be something I should expect to understand either.
many people are using chatbots as personal therapists, a friend to vent to, a diary that provides feedback, etc.
We don't feel sorry for this nonsense we punish it, whether it's "real" or not you have no right to share your fantasies of violence and real or not you need to be treated like they are and removed from society.
https://news.ycombinator.com/item?id=49961057
https://www.techspot.com/news/114091-florida-woman-used-clau...
This thread seems to be a dupe of that story.
An arrest isn't a conviction but this kind of fantasizing in any manner is probable cause for an arrest and grounds for a trial.
Arguably there should be a reporting duty for these kinds of thing.
I wish more people making threats got this treatment.
What if you get a peek at the diary of someone you actually know. You've know them for 20 years and they've never done anything violent in all that time - just the opposite in fact. The entry says they got a gun but you have no evidence of that other then the words in the diary. Do you call the police? Plot twist it's your Mom - do you still call the police?
That might be because the vast majority of people showing those flags are just regular people frustrated, not actual psychopaths. Rather than punishing this behavior, it should be treated to help them manage frustration.
Hell, emotional intelligence should be taught since kindergarten, it's not surprising we mostly behave like savages in clothes, we focused one half of the brain, and left the rest to chance.
Yes therapeutic intervention is ultimately how you stop these things from progressing but if you've gotten to the point that people see your threats and they seem plausible you need law enforcement and the judiciary involved deciding if you need to be removed from society not just a pat on the head from a social worker. Ideally you can get the help you need incarcerated.
At minimum this Florida woman needs to have her rights to own a firearm removed and be put on probation and in some kind of court ordered therapy.
Rights revoked and court punishment over fucking words. JFC.
There's a difference between making a threat and posing a threat. You and whoever gave you the ideology that's bouncing around inside your head are the problem more than some crazy typing garbage into chatgpt.
This is the part that law doesn't understand. It's binary, crime or no crime, threats or no threats, there is no gradient. The judge and jury can be the gradient, but as soon as precedence grows, that margin keeps shrinking.
One customer screams at another at a Wall Mart: "I will kill you for this!". How do we judge 'reasonably actionable' in this case? Words can't kill, so it's always an empty threat?
We all know that famous "I know it when I see it definition" but I think that's an exception rather then a rule. A lot of those things really can be given a reasonably clear definition.
People are putting a weird amount of moral weight on LLM prompts that they don't apply to other programs.
As for writing in a text format that was not ready by others or as fiction, I am thinking it has to do with intent.
Go a step further, she creates a bot that hits open chat bots with instructions on 'be a jerk'. Who is responsible?