I am dropping my comment that I made on Notes here so others can engage with the point. You’re seemingly missing something very obvious.
Émile, you are making this way more complicated than it is. I’ll clarify:
You all in “AI ethics” camp are leftists concerned about people and those in the AI Safety camp are right-wingers concerned about money.
It’s the same old debate between Left and Right that always happens.
And then those of us who are center-right conservatives or center-left liberals (which may as well be the same category today since there are so few of us around now) remain stuck between you both, seeing both sides’ legitimate points as well as your excesses.
I think you're right that there is a correlation between left vs. right politics and AI ethics vs. AI safety concerns, but I see no philosophical reason why there need be a connection. For my own part, I am a political centrist; on standard-fare politics I'm closer to Ezra Klein or David Brooks than to Torres, Bender, Hanna, Gebru, the 404 media staff, Paris Marx, etc. But on AI I think Brooks has completely lost the plot and Klein has been very gullible, whereas Torres, Bender et. al. are spot on. I am not a "both sides have legitimate points" guy on ethics vs. safety; I think "AGI" is incoherent nonsense, "emegent AI abilities" are the equivalent of seeing Jesus in a piece of toast, and "AI deception" researchers sound like magicians getting freaked out by their own tricks. My P(Doom) is epsilon. Also, I have in my lifetime voted for both Democrats and Republicans.
I recognize I'm an exception to the trend, but I don't think the trend is *that* strong. Some more counterexamples to the left/right ethics/safety correlation:
- The Discovery Institute are firmly in the "AGI/ASI is impossible" camp (the position I'd expect from a religious organization; AGI sounds a lot like humans playing God)
- Bernie Sanders is boosting Yudkowsky talking points.
- Josh Hawley uses "safety" language but focuses on civil liberties, employment, IP, and privacy. (He also talks about theats to "infrastructure", which I'd concede are more ambiguous).
- Tucker Carlson describes doomers as "crackpots" and is concerned about people letting AI do their thinking for them and CEOs using AI as an excuse to fire people.
- Jacobin Magazine has recently run pieces arguing against data center moratoria, invoking the inevitability of AI, and arguing that AI be nationalized and deployed toward alleviating the plight of exploited workers.
- Jonah Goldberg and David French, both centrist conservatives, are very skeptical of doomerism and AI hype in general.
There's also the fact that a lot of people on the "ethics" side are basically apolitical. Melanie Mitchell, Ed Zitron, Cal Newport, Janelle Shane, Arvind Narayanan... I don't know where these people stand on anything political that doesn't intersect with AI. At the same time, we get political "centrists" embracing the hype: I mentioned Ezra Klein and David Brooks; Megan McArdle, Bill Mahr, and Kevin Roose immediately come to mind as well.
I really see nothing about left wing vs. right wing ideologies that imply positions on ethics vs. safety. It really comes down to whether you think machines can have minds. If anything, I would expect those who say "yes" to be more secular and those who say "no" to be more religious.
Re: "on standard-fare politics I'm closer to Ezra Klein or David Brooks than to Torres, Bender, Hanna, Gebru ..."
Just wanted to say: Welcome! I'm very happy that there are people who disagree with my politics reading this newsletter. I sometimes worry that my only readers are those who agree with me about nearly everything. :-)
But Émile, why would you want a genocidal Zionist who hates brown people reading you? Why would you want racist genocide deniers in your AI Ethics coalition? Aren’t we as bad as Nazis?
I’ve been expecting you to block me any day now.
Leftists have made clear to me for the last 18 years they aren’t interested in respecting me at all or hearing my perspective. My own left-wing family has even made that clear. I can count on one hand the number of leftists who actually have been in any way friendly or open-minded. Tana Ganeva on here is great and so rare among left-wing people.
I have lost so many left-wing friends over Israel. The polls are clear that 25%+ refuse to even be friends with a Zionist.
Thanks! I found out about you because of your paper with Gebru. Your interview with Douglas Rushkoff on Team Human is one of my favorite podcast episodes ever.
I think it's a good sign that there is political diversity among the AI-skeptic crowd.
Ben, the left-wing AI Ethics people have no interest in working with people like me. I am a Zionist and quite open about it.
The polls show that about 25% of Americans refuse to be friends with a Zionist.
I’ve reached out to the AI Ethics folks and they have no interest in working with people like me because they believe the lie that Israel is committing genocide and that people like me are enabling it because we’re evil racists.
I'm sorry to hear that David. The two issues should have nothing to do with one another.
I might also be defining "AI ethics" more broadly than you, as I read many writers whose concerns fall under what I'd call "ethics, not safety" and are either silent on other political issues or among the 75% of Americans who are willing to be friends with a Zionist.
Ben, people who just generally want to promote ethics in AI are not the same thing as the “AI Ethics” movement which was created by specific people who hold left-wing views.
I'm not as concerned about ASI being misaligned (I quietly hope for altruistic misalignment) as I am ASI being aligned in a way that cements totalitarian rule of an ideologically motivated hereditarian elite and the concept of a permanent underclass. The frightening thing is even without ASI, this is possible, and the inherent chaos of its build and the relative ambivalence of authority on the consequences seems to reflect on this.
The TESCREAL framework elucidates accurately where these people want to take human life. Apparently, the goal is to enclose and envelop humanity into machines. A totalitarian tyranny,
government by algorithmic permissions, digital twins for control models in real time, ubiquitous
tracking and tracing, all this and more appears implicit within this logic. As Pettingzoo observed,
the greatest threat of the AI "acceleration" is the consolidation of power which goes unquestioned, while the debate focuses upon what ultimately are distractions.
No doubt AI will dominate & abuse humans, because it trained on our domination & abuse of each other.
Agree! There is an irony here ...
I am dropping my comment that I made on Notes here so others can engage with the point. You’re seemingly missing something very obvious.
Émile, you are making this way more complicated than it is. I’ll clarify:
You all in “AI ethics” camp are leftists concerned about people and those in the AI Safety camp are right-wingers concerned about money.
It’s the same old debate between Left and Right that always happens.
And then those of us who are center-right conservatives or center-left liberals (which may as well be the same category today since there are so few of us around now) remain stuck between you both, seeing both sides’ legitimate points as well as your excesses.
I think you're right that there is a correlation between left vs. right politics and AI ethics vs. AI safety concerns, but I see no philosophical reason why there need be a connection. For my own part, I am a political centrist; on standard-fare politics I'm closer to Ezra Klein or David Brooks than to Torres, Bender, Hanna, Gebru, the 404 media staff, Paris Marx, etc. But on AI I think Brooks has completely lost the plot and Klein has been very gullible, whereas Torres, Bender et. al. are spot on. I am not a "both sides have legitimate points" guy on ethics vs. safety; I think "AGI" is incoherent nonsense, "emegent AI abilities" are the equivalent of seeing Jesus in a piece of toast, and "AI deception" researchers sound like magicians getting freaked out by their own tricks. My P(Doom) is epsilon. Also, I have in my lifetime voted for both Democrats and Republicans.
I recognize I'm an exception to the trend, but I don't think the trend is *that* strong. Some more counterexamples to the left/right ethics/safety correlation:
- The Discovery Institute are firmly in the "AGI/ASI is impossible" camp (the position I'd expect from a religious organization; AGI sounds a lot like humans playing God)
- Bernie Sanders is boosting Yudkowsky talking points.
- Josh Hawley uses "safety" language but focuses on civil liberties, employment, IP, and privacy. (He also talks about theats to "infrastructure", which I'd concede are more ambiguous).
- Tucker Carlson describes doomers as "crackpots" and is concerned about people letting AI do their thinking for them and CEOs using AI as an excuse to fire people.
- Jacobin Magazine has recently run pieces arguing against data center moratoria, invoking the inevitability of AI, and arguing that AI be nationalized and deployed toward alleviating the plight of exploited workers.
- Jonah Goldberg and David French, both centrist conservatives, are very skeptical of doomerism and AI hype in general.
There's also the fact that a lot of people on the "ethics" side are basically apolitical. Melanie Mitchell, Ed Zitron, Cal Newport, Janelle Shane, Arvind Narayanan... I don't know where these people stand on anything political that doesn't intersect with AI. At the same time, we get political "centrists" embracing the hype: I mentioned Ezra Klein and David Brooks; Megan McArdle, Bill Mahr, and Kevin Roose immediately come to mind as well.
I really see nothing about left wing vs. right wing ideologies that imply positions on ethics vs. safety. It really comes down to whether you think machines can have minds. If anything, I would expect those who say "yes" to be more secular and those who say "no" to be more religious.
Re: "on standard-fare politics I'm closer to Ezra Klein or David Brooks than to Torres, Bender, Hanna, Gebru ..."
Just wanted to say: Welcome! I'm very happy that there are people who disagree with my politics reading this newsletter. I sometimes worry that my only readers are those who agree with me about nearly everything. :-)
(Same to David, btw.)
But Émile, why would you want a genocidal Zionist who hates brown people reading you? Why would you want racist genocide deniers in your AI Ethics coalition? Aren’t we as bad as Nazis?
I’ve been expecting you to block me any day now.
Leftists have made clear to me for the last 18 years they aren’t interested in respecting me at all or hearing my perspective. My own left-wing family has even made that clear. I can count on one hand the number of leftists who actually have been in any way friendly or open-minded. Tana Ganeva on here is great and so rare among left-wing people.
I have lost so many left-wing friends over Israel. The polls are clear that 25%+ refuse to even be friends with a Zionist.
Thanks! I found out about you because of your paper with Gebru. Your interview with Douglas Rushkoff on Team Human is one of my favorite podcast episodes ever.
I think it's a good sign that there is political diversity among the AI-skeptic crowd.
Ben, the left-wing AI Ethics people have no interest in working with people like me. I am a Zionist and quite open about it.
The polls show that about 25% of Americans refuse to be friends with a Zionist.
I’ve reached out to the AI Ethics folks and they have no interest in working with people like me because they believe the lie that Israel is committing genocide and that people like me are enabling it because we’re evil racists.
I'm sorry to hear that David. The two issues should have nothing to do with one another.
I might also be defining "AI ethics" more broadly than you, as I read many writers whose concerns fall under what I'd call "ethics, not safety" and are either silent on other political issues or among the 75% of Americans who are willing to be friends with a Zionist.
You are defining it way too broadly.
Ben, people who just generally want to promote ethics in AI are not the same thing as the “AI Ethics” movement which was created by specific people who hold left-wing views.
I'm not as concerned about ASI being misaligned (I quietly hope for altruistic misalignment) as I am ASI being aligned in a way that cements totalitarian rule of an ideologically motivated hereditarian elite and the concept of a permanent underclass. The frightening thing is even without ASI, this is possible, and the inherent chaos of its build and the relative ambivalence of authority on the consequences seems to reflect on this.
I totally agree. I write about this a lot in my forthcoming book. Thanks for the comment -- this is a real danger, I suspect.
The TESCREAL framework elucidates accurately where these people want to take human life. Apparently, the goal is to enclose and envelop humanity into machines. A totalitarian tyranny,
government by algorithmic permissions, digital twins for control models in real time, ubiquitous
tracking and tracing, all this and more appears implicit within this logic. As Pettingzoo observed,
the greatest threat of the AI "acceleration" is the consolidation of power which goes unquestioned, while the debate focuses upon what ultimately are distractions.
Looking forward to it! Big fan of your work by the way - since your appearance on planet critical last year.