Rethinking the purpose of AI with Stuart Russell
Mark Leonard's World in 30 Minutes
00:00
The Risks of Loss of Control
I'd like to keep separate the issue of lethal autonomous weapons, at least for the time being from this issue about control. The risk is not likely to be from killer robots in the near future because killer robots are probably going to be built with fairly simple-minded AI. But the risks coming from loss of control can be extremely subtle. And I think social media, again, is a warning. It happened over several years and we didn't realize that it was going on. To some extent, we have lost control because despite the efforts of governments, these algorithms are protected by trillion-dollar multinational corporations.
Transcript
Play full episode
