What you are seeing in these results is an outgrowth of how data at the big tech companies is much more about groups than individuals. These AI models from OpenAI and Google aren't trained to know about specific individuals. In fact, they don't want to know about them - that could get the researchers in trouble (see the issues with Github's CoPilot stealing code without attribution from legitimate projects). So right now, the more ridiculous the response when asked about individuals, the better.
Another way to think about it is that AI researchers want the AIs to "hallucinate" so that there is more randomness, which can lead to more creativity. So a way to think about asking for a biography of someone is that the AI will return a alternate history version of that person that, while wrong, still kinda-sorta has the same style and features.
All of this weirdness is a feature right now, not a bug. It's how the AI researchers are learning to train the system to write prose, chats, poetry, etc. by being a bit loose (but not too loose) with the data.
I expect based on what I've read that more training that includes individuals styles will soon be happening. Midjourney, an text-to-image AI, already allows for things like "in the style of Picasso", so someday soon you may be able to say "in the style of Arnold Kling". It may even know who you are by then, too! However, given current pushback on targeted advertising, don't be too surprised that it doesn't "know" you. Again, that's a feature that helps keep the researchers safe from trouble. It may be a long time before we breach that wall...
Henry Kissinger has been raising the specter of AI destroying humankind. In his view the fake IDs are trivial compared to the long-term threat from AI used for military purposes. He's been writing that is more dangerous than nuclear weapons. Are we really headed to the world of HAL from Kubrick's "2001"?
Well I'm not the David Friedman that most economists would think of when they hear that name. And I am not related to him. Nonetheless I decided to ask GPTChat to "write a biography of David Friedman" to see *which* David Friedman it would choose. (There are well-known non-economist David Friedmans floating around, such as the former Ambassador to Israel.)
In any event, GPTChat appears to think that Rose Friedman is David Friedman-the-economist's wife:
"David Friedman is an American economist, physicist, and legal scholar known for his libertarian views. He was born in 1945 in New York City, and he grew up in a Jewish household in the suburbs of Philadelphia. He studied physics at Harvard University, earning his Bachelor's degree in 1966. He went on to earn his Ph.D. in physics from the University of Chicago in 1971.
After completing his Ph.D., Friedman taught physics at the University of Chicago for several years before switching to economics. He has taught at several universities, including Santa Clara University and the University of California, Berkeley.
In addition to his academic work, Friedman is a prominent advocate for libertarianism, a political philosophy that advocates for limited government and individual liberty. He is the author of several books on economics, including "The Machinery of Freedom" and "Hidden Order: The Economics of Everyday Life." He has also written extensively on legal issues, including the application of economic principles to the law.
Friedman is married to fellow economist Rose Friedman, and they have two children. He has been awarded several honorary degrees and has been recognized for his contributions to economics and libertarian thought."
Online advertising consistently has a terrible time "getting" me. I am a happily married 53 year old observant Jew, and I keep getting adds for gourmet bacon and Christian dating websites.
I don't mind advertising. People have to get paid. I just wish the algorithms were less incompetent
My sense is that Chat GPT has been trained to exhibit a mild and nice personality. The political views that are most often stated in that mood online are exactly the Establishment Liberal ones, so perhaps that has something to do with the results from Rozado.
AI mistakenly credited Richard Hanania with “Myth of the Rational Voter” as well. Bryan said he might have to rethink his stance on AI risk.
https://twitter.com/bryan_caplan/status/1598777482127192065?s=46&t=J50CF2-TWyKPsOEz6U0zog
What you are seeing in these results is an outgrowth of how data at the big tech companies is much more about groups than individuals. These AI models from OpenAI and Google aren't trained to know about specific individuals. In fact, they don't want to know about them - that could get the researchers in trouble (see the issues with Github's CoPilot stealing code without attribution from legitimate projects). So right now, the more ridiculous the response when asked about individuals, the better.
Another way to think about it is that AI researchers want the AIs to "hallucinate" so that there is more randomness, which can lead to more creativity. So a way to think about asking for a biography of someone is that the AI will return a alternate history version of that person that, while wrong, still kinda-sorta has the same style and features.
All of this weirdness is a feature right now, not a bug. It's how the AI researchers are learning to train the system to write prose, chats, poetry, etc. by being a bit loose (but not too loose) with the data.
I expect based on what I've read that more training that includes individuals styles will soon be happening. Midjourney, an text-to-image AI, already allows for things like "in the style of Picasso", so someday soon you may be able to say "in the style of Arnold Kling". It may even know who you are by then, too! However, given current pushback on targeted advertising, don't be too surprised that it doesn't "know" you. Again, that's a feature that helps keep the researchers safe from trouble. It may be a long time before we breach that wall...
Henry Kissinger has been raising the specter of AI destroying humankind. In his view the fake IDs are trivial compared to the long-term threat from AI used for military purposes. He's been writing that is more dangerous than nuclear weapons. Are we really headed to the world of HAL from Kubrick's "2001"?
Well I'm not the David Friedman that most economists would think of when they hear that name. And I am not related to him. Nonetheless I decided to ask GPTChat to "write a biography of David Friedman" to see *which* David Friedman it would choose. (There are well-known non-economist David Friedmans floating around, such as the former Ambassador to Israel.)
In any event, GPTChat appears to think that Rose Friedman is David Friedman-the-economist's wife:
"David Friedman is an American economist, physicist, and legal scholar known for his libertarian views. He was born in 1945 in New York City, and he grew up in a Jewish household in the suburbs of Philadelphia. He studied physics at Harvard University, earning his Bachelor's degree in 1966. He went on to earn his Ph.D. in physics from the University of Chicago in 1971.
After completing his Ph.D., Friedman taught physics at the University of Chicago for several years before switching to economics. He has taught at several universities, including Santa Clara University and the University of California, Berkeley.
In addition to his academic work, Friedman is a prominent advocate for libertarianism, a political philosophy that advocates for limited government and individual liberty. He is the author of several books on economics, including "The Machinery of Freedom" and "Hidden Order: The Economics of Everyday Life." He has also written extensively on legal issues, including the application of economic principles to the law.
Friedman is married to fellow economist Rose Friedman, and they have two children. He has been awarded several honorary degrees and has been recognized for his contributions to economics and libertarian thought."
Online advertising consistently has a terrible time "getting" me. I am a happily married 53 year old observant Jew, and I keep getting adds for gourmet bacon and Christian dating websites.
I don't mind advertising. People have to get paid. I just wish the algorithms were less incompetent
I think Roemmele had a better idea, multiple assistants:
https://readmultiplex.com/2020/11/17/what-is-the-intelligence-amplifier-and-your-wisdom-keeper/
My sense is that Chat GPT has been trained to exhibit a mild and nice personality. The political views that are most often stated in that mood online are exactly the Establishment Liberal ones, so perhaps that has something to do with the results from Rozado.