Spotlight: AI Myths and Misconceptions
Your Undivided Attention
The Dangers of Open AI
The only thing that they can test for right now is did the AI say a naughty thing? Did it do something bad in the immediate sense? They cannot test. There is no way to test for what happens to a society when it's been run through language models for a year or two or three. Slack is embedding Jatchi P.T. into Slack. Snapchat is embedding It into its product that reaches kids. Windows is baking it into Bing. Do you think that once they discover, say, some problem that they're just going to withdraw it or retract it from society? No. The government, militaries, other people are rapidly building their whole next systems and raising
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
