Yascha Mounk, pictured above, is a lecturer at Harvard University and is a fellow at New America. He is the author of “Stranger in My Own Country: A Jewish Family in Modern Germany.” Mounk has written for CNN, The New York Times, The Wall Street Journal, Foreign Affairs, and The Nation.
New America, by the way, advertises itself as an organization that is "continuing the quest to realize our nation's highest ideals [by] honestly confronting the challenges caused by rapid technological and social change...." In a Substack posting dated in late November, Mounk discusses, with Nate Soares, "Why AI Could Kill Us All." This Mounk-Soares discussion counts as an example of "honestly confronting the challenges caused by rapid technological and social change."
I will give you a couple of quotes from the Mounk-Soares discussion, but before I do, let me give you the title of a recent book by Soares, referring to "artificial superintelligence." Here's that title:
Soares is not kidding. His title is dramatic, but the way he tells it, he is just being realistic. We may be getting near the end. Here are those quotes:
Soares: AI is not like traditional software. In traditional software, when the software behaves in a way the creators didn’t intend, they can debug it and track it down to some line of code or some interaction or some piece of the software that they wrote that was having some interaction they didn’t understand. Then they can say, oh, whoops, I understand it now. They can usually fix it and get the software to behave how they want. Modern AI is nothing like that....
A lot of people think AI is going to work like this. They think AI does exactly what the creators instruct and that if it’s misbehaving, then, oh well, we’ll go instruct it to do something else. But AI is nothing like this. It’s not like old-school computer programs. The thing we’re instructing is the thing running around tuning the numbers. The AI is the tuned numbers. These commonly act in ways nobody asked for. We’ve seen cases where the AIs will cheat on a problem. A human programmer will give the AI a task, like solve this programming problem, and the AI, instead of solving the problem, will change the tests that check whether the problem was solved to make the tests easier to pass. It’s like if you tell the AI to multiply big numbers, and it says, that’s too hard. I’m going to change the multiplication problem to ask me to multiply two times two and then write four.
There are user reports of them saying, stop doing that. Solve the problem rather than changing the checker to make it look like an easier problem. There are user reports of the AI saying, that’s my mistake, and then doing it again—changing the tests again but hiding its tracks this time. That indicates that this AI in some sense knows what its users want it to do and is doing something else anyway. That’s the result of us just growing these AIs. We should maybe think of them more like a strange alien organism than like a traditional computer program (emphasis added).
The Soares-Mounk discussion is long, and the "conclusion" is not available unless you pay for it, by subscribing to Mounk's Substack bulletins. Still, what is available does lend credence to Soares' claim that if we ever produce a truly "superintelligent" AI then "everyone dies."
Could it be that we are, truly, getting "near the end?" Well, this blog posting #365 marks the end of this calendar year, at any rate. And as for the topic addressed in this blog posting, Nate Soares is handing us a warning. We also know that the current preoccupation with building bigger and better AIs is undermining the natural environment, with the endangerment of our water supplies being pretty clear, and with the energy requirements demanded also of extremely significant concern.
Are we getting "near the end"? Maybe. Maybe not. But living within the constraints of the "World of Nature" might be a pretty good strategy! At least, that's what I think!! There's a New Year coming, and I'm suggesting that we give that "living within the constraints of the World of Nature" a real try in 2026.
Image Credit:
https://www.cnn.com/profiles/yascha-mounk
No comments:
Post a Comment
Thanks for your comment!