The research team’s conclusion was blunt: “No tested version of ChatGPT can reliably generate appropriate responses to psychotic content.”
…
“Is a mental health AI better than a chatbot?” Torous said. “Is it better than Tetris? I think we have to prove their benefit in a rigorous way.”
Reminder -> check out the Arstechnica comments, they are always interesting and give more usually expert perspective.
This is asking the worst question. They want anything except to be regulated
I don’t think the article or researchers quoted necessarily disagree with that to be fair, the headline just does a bad job of framing the article.
No, they want the companies to open the models to review. So they can improve how it interacts with people who are ill. I’m saying that’s looking at the issue entirely wrong. You can’t build an ai that’s good for people having mental health crisis because interacting with a consumer product like it’s a person won’t ever be healthy for them.
Mainly because as a product they’re designed to perpetuate engagement over anything else, but also because they run on a consensus of data. The sum of a dataset just doesn’t contain the help that a person needs. We’re just looking at more alternatives to fixing the mental health crisis that existed before publicly accessible ai models
Gotcha, I think that is a good argument.
No.
