Most people in college today believed, "If I graduate from college, I can get a white collar job that will be reasonably secure, pay reasonably well, and be more respectable than a blue collar job." They are scared s**tless that ai is taking that career path away from them. Since the heart of the Democratic Party is in academia, and ai threatens to take away some of the reason for schooling, of course it will be booed.
I agree but would add another factor to the job that most college-bound are looking for, especially women. Physical safety and some combination of clean, indoors, 9 to 5, and minimal physical effort.
FYI: “This post was inspired by Arnold Kling's post and comments from readers Roger Sweeny and Stu. It was conceived and written in collaboration with Claude, Anthropic's AI assistant.”
It seems that in general, anti forces on most issues, AI included, feel more strongly than Pro forces. They are motivated by anger, not by rational considerations. They will tend to show more in the election. If one believes (like I do), that the race for dominance between the US and China re AI is crucial to the fate of the Free World, and China allows far fewer Anti (anything) noise, the rise of the Left here poses real danger to US future.
“It seems that in general, anti forces on most issues, AI included, feel more strongly than Pro forces. They are motivated by anger, not by rational considerations.”
On individual issues, you are surely correct.
In terms of voting by the college-educated, I strongly suspect you are correct as well.
There is a lot of money in opposing development for any non-illegal reason. It doesn’t matter why or if the opposition is ultimately successful. For this reason it is not irrational.
How much of this divide can be attributed to simple majority rule, and could be avoided by requiring 2/3 to pass all legislation? I am serious.
Hysteresis is why thermostats start heating at the set point and stop heating when the temperature has risen one degree (or two, or a half, or whatever). If they tried to keep the temperature within .001 degree, they'd be inefficiently turning on and off every few seconds, or every time someone walked past and ruffled the air currents.
Passage by a simple majority means that changing just one or two seats can flip policies from Obama to Trump to Biden to Trump, or (somebody) to Youngkin to Spansberger. When they flip like that, the new party in power knows it has a good chance of being turned out next election, so they pass as much wild legislation as possible. Cram it through now, the hell with the consequences, because the future is unpredictable.
Whereas if it took 2/3 to pass all legislation, very few elections would result in wild swings from 70-30 to 30-70. Most would be from 55-45 to 45-55, or 60-40 to 40-60; legislators would know that the wild crazy bills would never garner enough support to pass. Legislators would have to work in gradual steps to get anything done.
Some elections would go from 60-40 to 70-30, and back to 65-35. But super majorities would not be the rule, unless districts were gerrymandered to the hilt. Maybe that would be one result, more than compromising to get things passed in gradual steps.
I believe Desantis has thrown some bones to the AI skeptics.
I’d also argue that to whatever extent it isn’t “salient”, it should be. Salient doesn’t mean taking an anti AI position. But you should 100 percent be thinking about what politicians think about this issue, even if that means explicitly favoring those who won’t call for more regulations etc.
Theres a reason this very blog now has a litany of posts (or cites posts) on AI related issues across a whole realm of areas (economy, education, future of work, warfare, risk of terror, etc etc etc). How should it not be salient?
AI is a threat to them because it introduces a great deal of uncertainty into their previously nearly-deterministic pathway to a cushy job and the “good life.” Like stu says in another comment on this post, they might have to get their hands dirty now.
The specter and reality of competition among countries for military power will limit regulation of AI in the U.S.
Supply-chain bottlenecks (power plants, data centers) and the politics of NIMBY will shape the geography of the AI industry in the U.S.. The general public good will be the face of NIMBY.
Frictions over public provision vs private entrepreneurship in (already highly regulated) energy production will color and hamper AI development.
Some of anti data centers/AI may be funded by those who want to hinder the US companies involved with AI, both hardware and software. Bueller? Bueller?
“…with small shifts in the up-for-grabs sliver of the white vote determining the outcome.”
I agree with everything else in this well-argued piece, both before and after this.
But in the most recent election, it was not just “sliver of the white vote” that moved away from the Omnicause party.
Blacks and Hispanics, and Asian-Americans, disproportionately but not solely male, also moved away from Dems in large numbers and caused not just a shift in the popular vote but a Dem loss across each of the swing states, and moved a few other states closer to being swing states.
Pew found Trump rose from 8% to 15% among black voters, 36% to 48% among Hispanic voters, and 30% to 40% among Asian voters.
Long live the Omnicause, I say.
It is our best chance against there being a President AOC or Mamdani.
What you may also see in the granular detail is that it is between property owners and property renters.
It seems even starker among those living in a city with strict controls on development, which increase the cost of building and thus the supply of rental units, driving rental prices up. The tenant is likely to consider themselves a victim and blame the greedy property owners and the dopy regulators. Push back invited.
“I would say that AI is only a salient issue among the hard core of college-educated Democrats.“
Not sure if the polls bear that out:
“Republicans and Democrats are now essentially tied on ‘more concerned than excited’ about AI (50% vs. 51%).”
“Asked whether AI’s impact on the US over the next 20 years will be positive or negative, only 17% of US adults said positive, while 35% said negative. Among AI experts, the picture is inverted: 56% said positive, 15% said negative. The gap between how the public and the people building AI see its future is one of the largest expert-public divides Pew has measured on any technology topic.”
Morevoer, AI has implications for nearly every salient public policy issue and it seems hard to separate it out as a separate issue. For example, polls show that Americans are excited about the potential for AI in health care, a top public policy issue of concern, but negative about its impact on the cost of living and the national debt. Construction of data centers has local effects on the price of real estate and housing. In many areas,data centers are occupying large swathes of space that would otherwise be developed for housing. Both retail electricity, and water rates have been rising faster than the overall rate of inflation and it is difficult if not flat out wrong, to suggest that AI will bring down the price of electricity or water. Although data center construction might provide a sugar high, long term data center employment doesn’t seem like the best of bets for a stable local economy:“More than half agree that they would oppose having a data center built in their community (55%), and at least three in five are worried about their energy usage (65%) and environmental impact (61%). Likewise, fewer Americans see an upside locally, as just 27% agree that data centers would significantly contribute to economic growth and job creation in their community.” (https://www.ipsos.com/en-us/latest-us-opinion-polls )
And then there is crime, for example, an issue of particular concern to the older population:
“Worry about internet fraud is widespread and intense, with AARP research indicating an average concern level of 7.1 out of 10. While 38% of adults report having personally experienced fraud, 58% express significant worry about becoming a victim, with older adults (50+) showing higher levels of concern than younger demographics.
Key factors driving this anxiety include:
High Personal Impact: 30% of adults cite becoming a victim as their primary fear, followed by 21% worrying about having no way to recover stolen funds.
Perceived Ubiquity: 87% of adults believe fraud can happen to anyone, yet 53% still harbor a contradictory belief that victims are somewhat blameworthy.
Rising Losses: Concern is fueled by a surge in theft, with losses jumping from $2.4 billion in 2019 to over $10 billion in 2023.“
Rightly or wrongly, AI is seen as contributing to this trend rather than as an effective counteragent.
Finally, Americans are deeply concerned about the national debt and its impact on the cost of living:
“75% of voters (including 71% of Democrats, 82% of independents and 75% of Republicans) say they would consider supporting a candidate from a political party they do not usually support, if that candidate had a clear plan to address the debt.” ( https://www.pgpf.org/press/2026-5-fci-press-release/)
Unfortunately AI has only been an excuse to drive up more government spending. GAO reports that agencies often fail to share lessons learned from AI acquisitions and that widespread, measurable budget savings from AI have not yet been fully realized or documented. (https://www.gao.gov/products/gao-25-107653 ). Given Congress’s failure to address the national debt and willingness to spend 20% of the federal budget annually on interest payments, it would probably be foolish to expect any amelioration of the debt crisis coming from AI. Rather, it is more likely that Congress will fall over itself coming up with more earmarks to benefit their clientele.
"As party elites differentiated ideologically through the 1970s and 80s and 90s, voters who held consistent ideological orientations started sorting into the party that matched their underlying views.
…And by 2024? The cleavage divides almost completely the cosmopolitans from the populists."
Unfortunately, the cosmopolitans lack a strong cultural grounding in lasting human values (tolerance and dependency programs and multiculturalism are all REACTIVE - they're not positive cultural values but are revisions of the older traditional memetic structure) and they're too fixated upon status. This caused them to veer into lunacy when confronted with reactive, radical ideas which sounded good (social justice ideology). This veering caused an elite split, which has been mostly unacknowledged but is still playing out right now.
Cheers Arnold. :) Here's my comments on your comments!
https://kylesaunders.substack.com/p/aipolitics-19-comments-on-arnold
Most people in college today believed, "If I graduate from college, I can get a white collar job that will be reasonably secure, pay reasonably well, and be more respectable than a blue collar job." They are scared s**tless that ai is taking that career path away from them. Since the heart of the Democratic Party is in academia, and ai threatens to take away some of the reason for schooling, of course it will be booed.
I agree but would add another factor to the job that most college-bound are looking for, especially women. Physical safety and some combination of clean, indoors, 9 to 5, and minimal physical effort.
That sounds like a pretty good description of most white collar jobs. As people used to say around here, "indoor job, no heavy lifting".
One of your best comments stu.
FYI: “This post was inspired by Arnold Kling's post and comments from readers Roger Sweeny and Stu. It was conceived and written in collaboration with Claude, Anthropic's AI assistant.”
https://scottgibb.substack.com/p/nobody-went-to-college-to-work-in
It seems that in general, anti forces on most issues, AI included, feel more strongly than Pro forces. They are motivated by anger, not by rational considerations. They will tend to show more in the election. If one believes (like I do), that the race for dominance between the US and China re AI is crucial to the fate of the Free World, and China allows far fewer Anti (anything) noise, the rise of the Left here poses real danger to US future.
“It seems that in general, anti forces on most issues, AI included, feel more strongly than Pro forces. They are motivated by anger, not by rational considerations.”
On individual issues, you are surely correct.
In terms of voting by the college-educated, I strongly suspect you are correct as well.
But in terms of voting overall, polling shows that most of the country - unlike we Substack commenters - voted *for* their preferred candidate rather than *against* their disfavored candidate. https://therightthingtodoandwhy.substack.com/p/for-and-against-the-2024-presidential
Say whatever else you want about so-called “MAGA” Republicans, but they are pro-America.
It is unclear whether or not that is true of the average Dem voter today.
And of course it is the opposite of the truth for the activist class in the Omnicause party.
How about a compromise- people vote for their favorite while extra motivated by their TDS? :)
There is a lot of money in opposing development for any non-illegal reason. It doesn’t matter why or if the opposition is ultimately successful. For this reason it is not irrational.
Not for the mob, just for the political leadership. At the end, the followers get screed so not so rational?
How much of this divide can be attributed to simple majority rule, and could be avoided by requiring 2/3 to pass all legislation? I am serious.
Hysteresis is why thermostats start heating at the set point and stop heating when the temperature has risen one degree (or two, or a half, or whatever). If they tried to keep the temperature within .001 degree, they'd be inefficiently turning on and off every few seconds, or every time someone walked past and ruffled the air currents.
Passage by a simple majority means that changing just one or two seats can flip policies from Obama to Trump to Biden to Trump, or (somebody) to Youngkin to Spansberger. When they flip like that, the new party in power knows it has a good chance of being turned out next election, so they pass as much wild legislation as possible. Cram it through now, the hell with the consequences, because the future is unpredictable.
Whereas if it took 2/3 to pass all legislation, very few elections would result in wild swings from 70-30 to 30-70. Most would be from 55-45 to 45-55, or 60-40 to 40-60; legislators would know that the wild crazy bills would never garner enough support to pass. Legislators would have to work in gradual steps to get anything done.
Some elections would go from 60-40 to 70-30, and back to 65-35. But super majorities would not be the rule, unless districts were gerrymandered to the hilt. Maybe that would be one result, more than compromising to get things passed in gradual steps.
I believe Desantis has thrown some bones to the AI skeptics.
I’d also argue that to whatever extent it isn’t “salient”, it should be. Salient doesn’t mean taking an anti AI position. But you should 100 percent be thinking about what politicians think about this issue, even if that means explicitly favoring those who won’t call for more regulations etc.
Theres a reason this very blog now has a litany of posts (or cites posts) on AI related issues across a whole realm of areas (economy, education, future of work, warfare, risk of terror, etc etc etc). How should it not be salient?
How can we really know whether people are voting on social identity versus political ideology? We probably can't.
Great post, Arnold — but what can the anti-AI folks actually do about it?
In the words of Seth Rollins, "Burn it down!"
I don't think that's going to happen, but a lot of people feel toward ai the way the audience feels toward the villain at the end of a revenge movie.
Democrats are the Party of Higher Education, AI is an existential threat to Higher Education, therefore Democrats will be the Party of Burn It Down.
AI is a threat to them because it introduces a great deal of uncertainty into their previously nearly-deterministic pathway to a cushy job and the “good life.” Like stu says in another comment on this post, they might have to get their hands dirty now.
Three predictions — each of uncertain magnitude:
The specter and reality of competition among countries for military power will limit regulation of AI in the U.S.
Supply-chain bottlenecks (power plants, data centers) and the politics of NIMBY will shape the geography of the AI industry in the U.S.. The general public good will be the face of NIMBY.
Frictions over public provision vs private entrepreneurship in (already highly regulated) energy production will color and hamper AI development.
Some of anti data centers/AI may be funded by those who want to hinder the US companies involved with AI, both hardware and software. Bueller? Bueller?
“…with small shifts in the up-for-grabs sliver of the white vote determining the outcome.”
I agree with everything else in this well-argued piece, both before and after this.
But in the most recent election, it was not just “sliver of the white vote” that moved away from the Omnicause party.
Blacks and Hispanics, and Asian-Americans, disproportionately but not solely male, also moved away from Dems in large numbers and caused not just a shift in the popular vote but a Dem loss across each of the swing states, and moved a few other states closer to being swing states.
Pew found Trump rose from 8% to 15% among black voters, 36% to 48% among Hispanic voters, and 30% to 40% among Asian voters.
Long live the Omnicause, I say.
It is our best chance against there being a President AOC or Mamdani.
What you may also see in the granular detail is that it is between property owners and property renters.
It seems even starker among those living in a city with strict controls on development, which increase the cost of building and thus the supply of rental units, driving rental prices up. The tenant is likely to consider themselves a victim and blame the greedy property owners and the dopy regulators. Push back invited.
“I would say that AI is only a salient issue among the hard core of college-educated Democrats.“
Not sure if the polls bear that out:
“Republicans and Democrats are now essentially tied on ‘more concerned than excited’ about AI (50% vs. 51%).”
“Asked whether AI’s impact on the US over the next 20 years will be positive or negative, only 17% of US adults said positive, while 35% said negative. Among AI experts, the picture is inverted: 56% said positive, 15% said negative. The gap between how the public and the people building AI see its future is one of the largest expert-public divides Pew has measured on any technology topic.”
(https://www.thealgorithmicbridge.com/p/how-america-turned-against-ai-according
Morevoer, AI has implications for nearly every salient public policy issue and it seems hard to separate it out as a separate issue. For example, polls show that Americans are excited about the potential for AI in health care, a top public policy issue of concern, but negative about its impact on the cost of living and the national debt. Construction of data centers has local effects on the price of real estate and housing. In many areas,data centers are occupying large swathes of space that would otherwise be developed for housing. Both retail electricity, and water rates have been rising faster than the overall rate of inflation and it is difficult if not flat out wrong, to suggest that AI will bring down the price of electricity or water. Although data center construction might provide a sugar high, long term data center employment doesn’t seem like the best of bets for a stable local economy:“More than half agree that they would oppose having a data center built in their community (55%), and at least three in five are worried about their energy usage (65%) and environmental impact (61%). Likewise, fewer Americans see an upside locally, as just 27% agree that data centers would significantly contribute to economic growth and job creation in their community.” (https://www.ipsos.com/en-us/latest-us-opinion-polls )
And then there is crime, for example, an issue of particular concern to the older population:
“Worry about internet fraud is widespread and intense, with AARP research indicating an average concern level of 7.1 out of 10. While 38% of adults report having personally experienced fraud, 58% express significant worry about becoming a victim, with older adults (50+) showing higher levels of concern than younger demographics.
Key factors driving this anxiety include:
High Personal Impact: 30% of adults cite becoming a victim as their primary fear, followed by 21% worrying about having no way to recover stolen funds.
Perceived Ubiquity: 87% of adults believe fraud can happen to anyone, yet 53% still harbor a contradictory belief that victims are somewhat blameworthy.
Rising Losses: Concern is fueled by a surge in theft, with losses jumping from $2.4 billion in 2019 to over $10 billion in 2023.“
Rightly or wrongly, AI is seen as contributing to this trend rather than as an effective counteragent.
Finally, Americans are deeply concerned about the national debt and its impact on the cost of living:
“75% of voters (including 71% of Democrats, 82% of independents and 75% of Republicans) say they would consider supporting a candidate from a political party they do not usually support, if that candidate had a clear plan to address the debt.” ( https://www.pgpf.org/press/2026-5-fci-press-release/)
Unfortunately AI has only been an excuse to drive up more government spending. GAO reports that agencies often fail to share lessons learned from AI acquisitions and that widespread, measurable budget savings from AI have not yet been fully realized or documented. (https://www.gao.gov/products/gao-25-107653 ). Given Congress’s failure to address the national debt and willingness to spend 20% of the federal budget annually on interest payments, it would probably be foolish to expect any amelioration of the debt crisis coming from AI. Rather, it is more likely that Congress will fall over itself coming up with more earmarks to benefit their clientele.
"As party elites differentiated ideologically through the 1970s and 80s and 90s, voters who held consistent ideological orientations started sorting into the party that matched their underlying views.
…And by 2024? The cleavage divides almost completely the cosmopolitans from the populists."
Unfortunately, the cosmopolitans lack a strong cultural grounding in lasting human values (tolerance and dependency programs and multiculturalism are all REACTIVE - they're not positive cultural values but are revisions of the older traditional memetic structure) and they're too fixated upon status. This caused them to veer into lunacy when confronted with reactive, radical ideas which sounded good (social justice ideology). This veering caused an elite split, which has been mostly unacknowledged but is still playing out right now.
https://jmpolemic.substack.com/p/the-schism-of-the-elites