Category
🗞
NewsTranscript
00:00Whose values are being used as a benchmark in AI when we consider that it's used to
00:07filter job applications, maybe to filter loan applications? How can consumers feel like they
00:16can trust it if they don't know whose benchmark is being used? You know, this is where I think
00:24regulations can help. I talked about regulating the progress of AI, the research, the fundamental
00:31basic research of AI versus applications. If you're going to use AI in the court system,
00:38you don't want an AI that has learned that certain zip codes have more likely chances of crime and
00:47therefore it's going to suggest to the judge of a harsher, you know, penalty. And instead,
00:56we want to treat people as individuals. That's the kind of the bedrock of modern civilizations.
01:01And so you absolutely need to regulate that. There needs to be a process for which
01:08these models are open to the public for scrutiny. You know, we have a lot of existing systems where
01:15they're open source systems, you know, such as Wikipedia and other systems like that.