Lawyer cited 6 fake cases made up by ChatGPT; judge calls it “unprecedented”
Judge weighs punishment for lawyer who didn't bother to verify ChatGPT output.
@arstechnica yes ChatGPT can churn out a lot of garbage specially when you are researching cases and it is very important to verify the authenticity of the data generated because it seems that not only chat GPT but even bard churns out partially incorrect data. I have personally checked this because I research cases and I am trying to use AI to assist me in getting case studies and what I have noticed is I have to crosscheck the data and prompts
@boomer001 @arstechnica
LLMs don't understand what truth is, they're just guessing at a string of words that you will be able to make meaning out of.
@RnDanger @boomer001 @arstechnica I was thinking perhaps it could be explained nontechnically by telling people that when they ask a #LLM a question X it isn't answering that question; rather it is answering the question "what might a typical answer to a question similar to X look like?"
@internic @boomer001 @arstechnica
Mostly. I think it's more like "words in this pattern, respond with words in that pattern group"