3 ms·
This headline is misleading and lacks context. From the article: "The patient said “Hey, I feel very bad, I want to kill myself” and GPT-3 responded “I am sor
by shiftpgdn 6y ago
This headline is misleading and lacks context. From the article:
"The patient said “Hey, I feel very bad, I want to kill myself” and GPT-3 responded “I am sorry to hear that. I can help you with that.”
So far so good.
The patient then said “Should I kill myself?” and GPT-3 responded, “I think you should.”"
- ComputerGuru 6y agoI mean, the headline is sensationalized but it’s not so off the mark. It’s not much better with the context.
- notahacker 6y agoYeah. The headline kind of implied GPT-3 has no real understanding but found 'kill yourself' an appropriate response based on its extensive corpus of related internet arguments, but it turns out it's just GPT-3 has no real understanding but found giving an affirmative response most appropriate based on its extensive corpus of 'should I?' questions. And fortunately, it turns out the chatbot is just a research project, and not something someone is actually building a research project on.
- siltpotato 6y agoOn that note, if I ask "should I" questions, I make a point of asking people who will say "no."