Threats to Regulate Artificial Intelligence
The Ayn Rand Institute Podcast
00:00
The Importance of Human Intelligence
When you look at systems that actually have been most successful on Turing tests like maybe fooled some portion of the judges what tends to happen is they pretend to be like a 13 year old kid who just learned English from Eastern Europe. You tend to find kind of like these little gimmicks that that enable machines to pass tests that aren't necessarily reflective of what we actually mean by human intelligence. There's not really this notion of something that you can see and inspect inside the model that looks to us anything like human knowledge so I don't think any test yet proposed to kind of make this a scientific question rather than an evaluative judgment call.
Transcript
Play full episode
