Why Public Disputes About AI Have Been Mostly Non-Partisan. So Far.

Generative AI - headlined by ChatGPT - is new and no one yet seems advantaged or disadvantaged by it as Republicans perceived they were by Covid and its vaccines. 

Both parties have reasons for wanting to keep big tech and its tools in check so there is general agreement on who the enemy is, if for very different reasons having to do with how control affects reputations and voting patterns. But since almost everything in American life these days eventually becomes a red versus blue issue, AI is probably not far behind - and the first attack is likely  to be someone accusing it of being woke, whatever that may mean.  JL

Jacob Stern reports in The Atlantic:

AI has so far managed to remain strikingly nonpartisan. You know something strange is afoot when Elon Musk comes out in favor of tech regulation. Or Kevin McCarthy and a left-wing Joe Biden appointee agree that one particular issue is a priority. Generative AI is new. People have not had time to take positions on the technology, let alone calcify them. Nor have Americans gotten many clear signals from their political leaders about how they should feel. “The different parties have different motivations, but they both have reasons to keep the tech industry in check.” (But) our political system excels at polarizing even the most anodyne issues. I wouldn’t bet against it.

You know something strange is afoot when Elon Musk comes out in favor of tech regulation. Or when Kevin McCarthy and a left-wing Joe Biden appointee agree that one particular issue is a priority. These are not people who tend to agree on, well, anything. But such are the nascent, topsy-turvy politics of artificial intelligence.

AI is not really a single issue you can be for or against the way you can with, say, guns or abortion. It is, to name just a few aspects, an economic issue (Will it replace millions of jobs?), a foreign-policy issue (What if China surpasses us?), and a political issue (Is it about to supercharge our misinformation problem?). But there is one overarching debate about AI: between the techno-utopians who think it will usher in a new age of prosperity and the techno-pessimists who think it will be a destructive, destabilizing force that might just usher in the end of the world. Those techno-utopians and techno-pessimists don’t break along strict political lines, though. Somehow, in an era when kitchen appliances and children’s books can erupt into new fronts in the culture war virtually overnight, AI has so far managed to remain strikingly nonpartisan.

The unpolarized status quo is due in part to the fact that generative AI is so new—or at least newly mainstream. AI fever began in earnest less than six months ago with the launch of ChatGPT. People simply have not had much time to take positions on the technology, let alone calcify them. Nor have Americans gotten many clear signals from their political leaders about how they should feel. “The main impetus for politicization around things is what the leaders are saying about it,” Lilliana Mason, a political scientist at Johns Hopkins University who studies polarization, told me. For example, COVID was not a political issue when it first entered the public consciousness in late 2019 and early 2020—but over time, Republican rhetoric downplaying the virus helped open a partisan gap.

Read: Return of the people machine

That hasn’t yet happened with AI. The Biden administration’s response to ChatGPT has largely amounted to a vague document, released in October, described as a “blueprint for an AI bill of rights.” In late March, when a reporter asked the White House press secretary about whether AI poses an existential risk, she laughed the question off. Last week, the president offered some not-particularly-revelatory remarks of his own. Several bills regulating AI have been introduced in Congress, but none has gone anywhere yet. Ted Lieu and Don Beyer, both Democratic members of the House Artificial Intelligence Caucus, told me that the issue has not yet become polarized.

The bipartisan honeymoon might come to an end when Biden or Donald Trump or even major media personalities such as Tucker Carlson and Rachel Maddow start staking out positions on the promise or menace of AI. Polarization could also arrive if AI starts victimizing large numbers of people, Mason said, whether that’s by taking jobs, abusing users, or rendering certain ways of life obsolete. (Some of this may already be happening to a few users.)

How exactly the parties would align—pro-AI or anti-AI or somewhere in between—is unclear. The possibilities are downright kaleidoscopic. Democrats, traditionally more concerned about long-term threats such as climate change and pandemics, might oppose AI development on the grounds of existential risk. Republicans, traditionally more concerned about the preservation of existing social structures and mores, might oppose it as a potentially destabilizing force. Or maybe they’d support it, given their distaste for government regulation. Then again, they might oppose it on the basis of their science-skepticism, whereas Democrats might support it because they’re “following the science.” Maybe both parties will oppose regulation for fear of ceding AI dominance to China.

Or maybe they’ll find common cause in their commitment to privacy and wariness of Big Tech. “There’s actually a lot of ground to agree on,” Deborah Raji, an expert on AI bias at the Mozilla Foundation, told me. “The different parties have different motivations, but they both have reasons to keep the tech industry in check.” Nevertheless, those differing motivations could become the fault lines along which factions split off. Already, debates have flared up on questions of bias and censorship, with liberals worried about ChatGPT’s propensity for racial prejudice and conservatives fixated on its refusal to utter racial slurs.

Even if matters of principle don’t turn AI partisan, money might. AI moguls have come out in favor of regulation, but it’s one thing to say that in the abstract and quite another to support actual legislation. Beyer worries about a scenario in which the AI industry, in an effort to fight bipartisan regulation efforts, dumps boatloads of cash into the coffers of one party or the other. It would not be the first industry to do so.

Part of what makes the politics of AI so tricky to get ahold of is that AI is an everything issue. It’s like so many different things—nuclear weapons, gain-of-function research, electricity—and in that sense not quite like any of them, which makes it hard to slot into any existing partisan framework. “If you asked 10 different people in Congress to define artificial intelligence,” Beyer told me, “you’d get at least 10 different answers.” It’s tough to split into two distinct camps when no one understands what you’re arguing about.

Read: AI is like … nuclear weapons?

If anyone can do it, though, Americans can. Our political system is dysfunctional in countless ways, but it excels at polarizing even the most anodyne issues. I wouldn’t bet against it.