The politics of AI, part two

This article is an on-site version of our Swamp Notes newsletter. Premium subscribers can sign up to get the newsletter delivered every Monday and Friday. Standard subscribers can upgrade to Premium here, or explore all FT newsletters

Hello Swampians, it’s Jonathan Derbyshire here, the FT’s US opinion editor, standing in for Rana.

Last week the Republican Party held a gathering in Dallas with an eye firmly on November’s midterm elections. This was not a convention, the New York Times cautioned, since President Donald Trump had insisted on stripping out “all the perfunctory, yet necessary, parts of a typical presidential convention and [keeping] the made-for-TV elements”. And principal among these, of course, was the president’s address to the GOP and Maga faithful, in which he promised every adult American a $5,000 “dividend” if the Republicans win the midterms. (It was later calculated that this stunt would cost somewhere in the region of $1.3tn. Tackled about this, vice-president JD Vance limply suggested tariff revenues would help to cover that eye-watering sum.)

Trump also, to the surprise of precisely no one, told the crowd to “pretend that I’m on the ballot” in November. I wrote in a Swamp Note last month that the Republicans’ predicament is that the “party is now entirely in the iron grip of a president whose approval ratings threaten to be a drag on its wider electoral chances”. There’s little to add to that analysis here.

Much the most interesting response to events in Dallas came from my colleague Ed Luce, in Friday’s edition of Swamp Notes. Ed pointed out that the Republican powwow was immediately preceded by outgoing Anthropic employee Jacob Coxon’s blood-curdling warning that his former colleagues “earnestly believe” that artificial intelligence “could kill us all by the end of the decade”. This apocalyptic outlook was corroborated by Anthropic’s Evan Hubinger, who said: “Jacob is correct here — we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade.”

The astonishing thing about the Dallas meeting, Ed wrote, was that no one, from the president down, thought to mention these warnings — despite the emergence of what Ed, in a column last month, has called America’s new AI-phobic consensus.

Until now, most of the political focus on AI has been directed towards energy-sucking data centres, opposition to the building of which is definitely bipartisan. (Rana Foroohar wrote about the “politics of data centres” here.) But what is rapidly becoming clear is that the attention of voters and lawmakers alike is now turning to the AI models themselves.

The week before Coxon broke cover, Bernie Sanders, the independent leftwing senator from Vermont, together with Rep. Greg Casar of Texas (a Democrat), launched the Ban Artificial Superintelligence Act. The proposed legislation would “permanently ban the development and deployment of superintelligent AI and temporarily pause advanced AI deployment until a federal regulator has established safety rules.”

One likely effect of Sanders’ move, irrespective of whether or not the legislation makes any Congressional headway, will be to sharpen the tensions inside the AI industry that Ed identified in his column. Market rivalry is intense, he noted, and gets expressed in the differing stances that firms take on regulation. “The AI companies that are performing best — notably Anthropic, which is preparing for AI’s largest-ever initial public offering — are in favour of regulation and closed-weight models, or those with private parameters.” The laggards, on the other hand, including Mark Zuckerberg’s Meta, “have found religion in open-weight models”.

Other lawmakers rushed to raise the alarm after Coxon’s remarks. Democratic senator Chris Murphy said AI companies were in a “blind race to build a death machine first”. His Republican counterpart Ted Cruz called the risks from AI models going rogue “dangerous and frightening”.

Most intriguing of all was a conservative gathering that also took place in Dallas on Wednesday, just as Trump was about to take the stage. Ben Smith of Semafor reports that Joe Allen, formerly the “transhumanism editor” on Steve Bannon’s War Room podcast, told a “hard-right crowd”: “You look at the data centre, this incubator for algorithmic immigrants — they’ve already broken out and invaded other companies. These immigrants are out of control.”

That’s one way to get a rightwing audience’s blood up about this stuff, I suppose. Whether it’s a line we’ll hear from Republican candidates for Congress is another matter. But what is undeniable is that AI is well and truly on the agenda in these midterms.

I’m delighted to call on my colleague Richard Waters as my respondent this week. Richard, you’ve probably forgotten more about the AI industry than I’ll ever know. How do you read this growing political disquiet, not just about data centres but about the apparently existential threat posed by the models themselves? And how do you think the big AI companies are going to respond? Can we expect even more Silicon Valley cash to pour into races across the country?

Recommended reading

• In the New York of Review of Books, Omer Bartov, a professor of Holocaust and genocide studies at Brown, reviews On Antisemitism: A Word in History by Mark Mazower. “A growing range of scholars and other observers seem to agree with Mazower that even if ‘some anti-Zionism masks anti-Jewish sentiments, much of it does not,’” Bartov writes. “But most major American Jewish organisations have stayed attached to the assumption that the main source of antisemitism in the world today is the anti-Israel left.”
• For the FT, Constanze Stelzenmüller examines, with some foreboding, the seismic implications of the massive surge of the far-right Alternative for Germany (AfD) in state elections in Saxony-Anhalt last Sunday. While the AfD fell short of an outright majority, its opponents ought not to breathe easy, Constanze cautions: “This vote, and the ensuing helpless wittering by democratic political forces in Berlin, signals far greater challenges ahead.” FT contributing editor Adam Tooze has also written about the AfD, in his Chartbook newsletter. He suggests that part of the AfD’s appeal lies in its ability to mobilise profound popular dissatisfaction with the German state. “The AfD’s fixation on migration and the cultural battle for Germany’s ‘soul’,” Adam writes, speaks to that sense of malaise in a way “that none of the establishment parties currently can”.
• The far right has been mobilising in Britain too in recent weeks — though on the streets rather than at the ballot box. In the London Review of Books, Daniel Trilling argues that “one of the main differences between today’s far right and the fascism of the mid-20th century is the absence of a uniformed paramilitary street wing”. Anti-migrant protests by squadrons of black-clad masked men in Dover and Portsmouth mark, he says, are “an attempt to create one — or at least to project the image of one”.

In the New York of Review of Books, Omer Bartov, a professor of Holocaust and genocide studies at Brown, reviews On Antisemitism: A Word in History by Mark Mazower. “A growing range of scholars and other observers seem to agree with Mazower that even if ‘some anti-Zionism masks anti-Jewish sentiments, much of it does not,’” Bartov writes. “But most major American Jewish organisations have stayed attached to the assumption that the main source of antisemitism in the world today is the anti-Israel left.”

For the FT, Constanze Stelzenmüller examines, with some foreboding, the seismic implications of the massive surge of the far-right Alternative for Germany (AfD) in state elections in Saxony-Anhalt last Sunday. While the AfD fell short of an outright majority, its opponents ought not to breathe easy, Constanze cautions: “This vote, and the ensuing helpless wittering by democratic political forces in Berlin, signals far greater challenges ahead.” FT contributing editor Adam Tooze has also written about the AfD, in his Chartbook newsletter. He suggests that part of the AfD’s appeal lies in its ability to mobilise profound popular dissatisfaction with the German state. “The AfD’s fixation on migration and the cultural battle for Germany’s ‘soul’,” Adam writes, speaks to that sense of malaise in a way “that none of the establishment parties currently can”.

The far right has been mobilising in Britain too in recent weeks — though on the streets rather than at the ballot box. In the London Review of Books, Daniel Trilling argues that “one of the main differences between today’s far right and the fascism of the mid-20th century is the absence of a uniformed paramilitary street wing”. Anti-migrant protests by squadrons of black-clad masked men in Dover and Portsmouth mark, he says, are “an attempt to create one — or at least to project the image of one”.

Richard Waters replies

Jonathan, I love the idea of immigration control for robots. People in tech already talk about whether to tax robots or give legal personhood to AI, so why shouldn’t their immigration status be next?

But daft as the idea is, it does serve to make an important point. People have a lot of inchoate worries about AI — how do politicians connect those to issues that have real political leverage? I don’t see it happening in this election cycle in the US, but I definitely think it’s coming.

To take the issues you raise. It was really interesting how the existential risks from AI became such a big topic last week, everyone was talking about it. People in AI have been warning about this for years, I first wrote a big piece about it 12 years ago (yes, that really dates me). I never quite understand why something suddenly breaks through and touches a public nerve, but it feels important.

However, much as Sanders wants to make this an immediate political issue, the general reaction isn’t outrage, it’s bewilderment. When something sounds like science fiction and even a lot of techies say it’s nonsense (like Nvidia CEO Jensen Huang), I don’t see how it gets political traction. People have plenty of other, more immediate problems to worry about.

Data centres have definitely become political, but to me this seems entirely about local development, not the beginning of a rebellion against the machines. The politicians who might have been vulnerable have generally inoculated themselves by talking tougher about development. Top prize for political cowardice goes to governor Greg Abbott of Texas, who last year said his state was going to be the “epicentre” of data centre development and has now ordered a pause on connecting them to the electricity grid.

I think AI will feed into politics when its economic effects become clearer. The concentrations of wealth and power from the technology are already stark, but we’re only at the very beginning of this shift. At some point, AI could really cut into jobs or depress earnings. How this will play out is hard to predict in advance. What will the impact be if large numbers of lawyers lose their jobs? Washington is full of lawyers, it wouldn’t feel as remote as coal miners in West Virginia. But I think we’ll have to wait an election cycle or two for something like that.

Your feedback

We’d love to hear from you. You can email the team on [email protected], contact Ed on [email protected] and Rana on [email protected], and follow them on X at @RanaForoohar and @EdwardGLuce. We may feature an excerpt of your response in the next newsletter

Recommended newsletters for you

The AI Shift — John Burn-Murdoch and Sarah O’Connor dive into how AI is transforming the world of work. Sign up

Unhedged — Robert Armstrong dissects the most important market trends and discusses how Wall Street’s best minds respond to them. Sign up