Arvind Narayanan

I can't give you a number saying, you know, 90% of the time it's going to be right. It just really varies depending on one's use case. So, I think Each of us has to put a little bit of trial and error into adapting it for our own specific purposes.

I can't give you a number saying, you know, 90% of the time it's going to be right. It just really varies depending on one's use case. So, I think Each of us has to put a little bit of trial and error into adapting it for our own specific purposes.

I can't give you a number saying, you know, 90% of the time it's going to be right. It just really varies depending on one's use case. So, I think Each of us has to put a little bit of trial and error into adapting it for our own specific purposes.

Criminal justice, for instance, right? So I don't think we should be making decisions about people based on these crude statistical formulas with some caveats, like I was saying earlier. If it's the judge who is empowered to make that decision, that's a different story. There is a lot of the same coil in hiring.

Criminal justice, for instance, right? So I don't think we should be making decisions about people based on these crude statistical formulas with some caveats, like I was saying earlier. If it's the judge who is empowered to make that decision, that's a different story. There is a lot of the same coil in hiring.

Criminal justice, for instance, right? So I don't think we should be making decisions about people based on these crude statistical formulas with some caveats, like I was saying earlier. If it's the judge who is empowered to make that decision, that's a different story. There is a lot of the same coil in hiring.

There are companies that claim that by analyzing a 30-second video of a candidate, of a job candidate, not even talking about their skills for the job, but about their hobbies or whatever, that they can do video analysis and look at the candidate's facial expressions and body language and that sort of thing and use that to derive a personality score, which companies should do their hiring based on.

There are companies that claim that by analyzing a 30-second video of a candidate, of a job candidate, not even talking about their skills for the job, but about their hobbies or whatever, that they can do video analysis and look at the candidate's facial expressions and body language and that sort of thing and use that to derive a personality score, which companies should do their hiring based on.

There are companies that claim that by analyzing a 30-second video of a candidate, of a job candidate, not even talking about their skills for the job, but about their hobbies or whatever, that they can do video analysis and look at the candidate's facial expressions and body language and that sort of thing and use that to derive a personality score, which companies should do their hiring based on.

There's so much more. There is AI for detecting which students in a school or college might be at risk of suicide or mental health difficulties. There have been investigations of all these kinds of AI tools and they barely work better than the flip of a coin. So I think these are the kinds of things we should be very suspicious of.

There's so much more. There is AI for detecting which students in a school or college might be at risk of suicide or mental health difficulties. There have been investigations of all these kinds of AI tools and they barely work better than the flip of a coin. So I think these are the kinds of things we should be very suspicious of.

There's so much more. There is AI for detecting which students in a school or college might be at risk of suicide or mental health difficulties. There have been investigations of all these kinds of AI tools and they barely work better than the flip of a coin. So I think these are the kinds of things we should be very suspicious of.

And unfortunately, these are the kinds of things that are often used in order to make very high stakes decisions about people.

And unfortunately, these are the kinds of things that are often used in order to make very high stakes decisions about people.

And unfortunately, these are the kinds of things that are often used in order to make very high stakes decisions about people.

For ChatGPT, yes. But here's the difference between ChatGPT and trying to predict if someone will commit a crime. You know, ChatGPT is just trying to do things like, you know, a typical thing you might use ChatGPT for is to translate text from one language to another, right? That's not like a fundamentally impossible task. It's something that humans can do.

For ChatGPT, yes. But here's the difference between ChatGPT and trying to predict if someone will commit a crime. You know, ChatGPT is just trying to do things like, you know, a typical thing you might use ChatGPT for is to translate text from one language to another, right? That's not like a fundamentally impossible task. It's something that humans can do.

For ChatGPT, yes. But here's the difference between ChatGPT and trying to predict if someone will commit a crime. You know, ChatGPT is just trying to do things like, you know, a typical thing you might use ChatGPT for is to translate text from one language to another, right? That's not like a fundamentally impossible task. It's something that humans can do.

And AI over time is learning to do it better, right? Or write code or whatever it is. On the other hand... Predicting what's going to happen in the future, no one knows. The universe doesn't know. It doesn't matter how much data you can throw at it. What we're seeing is that these technologies are not really getting better. They haven't got better in decades.

And AI over time is learning to do it better, right? Or write code or whatever it is. On the other hand... Predicting what's going to happen in the future, no one knows. The universe doesn't know. It doesn't matter how much data you can throw at it. What we're seeing is that these technologies are not really getting better. They haven't got better in decades.