#64 – Michael Aird on Strategies for Reducing AI Existential Risk
Hear This Idea
00:00
The Importance of Interpretability in AI Takeover
There's going to be a lot of other problems that just don't connect very much so for example the way to deal with like bias and fairness type stuff um I mean there's lots of ways but it could. One problem is like toxic speech from right like polarization they could change the type stuff yeah recommend their algorithmsYeah And there are some ways this does connect to the AI eccentric thing or like some ways that the same intervention is helpful for both. So one thing that's helpful for both is interpretability at the moment we have these systems that are like very very very large.
Transcript
Play full episode
