News for hackers. Stuff that matters.

A comment on Medical chatbot using OpenAI’s GPT-3 told a fake patient to kill themselves. Parent comment

Re:Medical chatbot using OpenAI’s GPT-3 told a fake patient ... by 908B64B197 on Friday February 26, 2021 @06:30PM (#26281084)
I didn't know it was made for a joke conference!

But it does illustrates how it's easy to get trapped into a local minima.

[ Parent | on HN ]
Re:Medical chatbot using OpenAI’s GPT-3 told a fake patient ... by wizzwizz4 on Friday February 26, 2021 @06:42PM (#26281177)
They're doing what you tell them to, not what you want them to. Same with GPT-3; it's predicting what (from its experience) it thinks a human would say, not what is ethical to say.
[ Parent | on HN ]
Re:Medical chatbot using OpenAI’s GPT-3 told a fake patient ... by rmrfrmrf on Friday February 26, 2021 @09:02PM (#26282000)
unless...
[ Parent | on HN ]