AI Phantom Menace Existential Dread Isn’t Yet a Political Weapon

1. In Dallas last week, the Republican Party gathered for a midterm-focused rally that doubled as a tribute to President Donald Trump’s enduring grip on the GOP. Amidst promises of $5,000 dividends and exhortations to “pretend I’m on the ballot,” a starkly different narrative was unfolding just outside the spotlight. Jacob Coxon, a former employee of AI safety firm Anthropic, had recently issued a chilling warning: his former colleagues “earnestly believe” that AI could kill all humans by the end of the decade, with some estimating the probability at greater than 10%.

Yet, as FT’s Jonathan Derbyshire noted in his latest Swamp Notes, not a single speaker at the Dallas rally mentioned these existential threats. This silence highlights a growing disconnect between the intense internal debates within the AI industry and the current priorities of American voters and politicians. While an "AI-phobic consensus" may be emerging among elites, it has yet to crystallize into a coherent political platform.

Data Centers to Doomsday

Until recently, political opposition to AI in the US has been largely NIMBYist (Not In My Backyard). Bipartisan outrage has focused on the physical footprint of the technology: energy-guzzling data centers straining local power grids and water supplies. Governors like Texas’s Greg Abbott, who once touted his state as the “epicentre” of AI development, have since pivoted to pausing grid connections, reflecting a pragmatic response to constituent anger over infrastructure strain rather than ideological opposition to the technology itself.

However, the conversation is shifting. The emergence of voices like Coxon’s, corroborated by Anthropic’s Evan Hubinger, has brought the concept of "existential risk" into the mainstream media cycle. This has prompted legislative reactions, such as Senator Bernie Sanders’ proposed Ban Artificial Superintelligence Act, which seeks to pause advanced AI deployment until federal safety rules are established. Even Republicans like Ted Cruz have joined the chorus, labeling rogue AI models as “dangerous and frightening.”

But as FT technology editor Richard Waters points out, translating this bewilderment into political traction is difficult. When the threat sounds like science fiction and is dismissed by industry titans like Nvidia’s Jensen Huang as nonsense it struggles to compete with immediate concerns like inflation, housing, and healthcare. “People have plenty of other, more immediate problems to worry about,” Waters observes.

The Industry’s Regulatory Split

The political debate is also being shaped by fierce competition within the AI industry itself. Waters notes a clear divergence in regulatory stances:

• The Leaders (e.g., Anthropic): Favor strict regulation and closed-weight models (private parameters), positioning themselves as responsible stewards of dangerous technology. This stance also serves to raise barriers to entry for competitors.

• The Laggards (e.g., Meta): Have embraced open-weight models, arguing that transparency and widespread access are safer than concentrating power in a few hands.

This corporate rivalry ensures that both sides will pour money into political lobbying, but their messages may cancel each other out, leaving voters confused rather than mobilized.

The Right-Wing Narrative: "Algorithmic Immigrants"

Perhaps the most intriguing development is how the far right is framing the AI threat. At a conservative gathering in Dallas, Joe Allen, a former editor for Steve Bannon’s War Room, described AI models as “algorithmic immigrants” that have “broken out and invaded other companies.” By framing AI through the lens of immigration and cultural invasion, right-wing populists are attempting to tap into existing anxieties about loss of control and national identity.

While this rhetoric may energize base voters, it remains unclear whether it will resonate with the broader electorate. As Waters suggests, the idea of “immigration control for robots” is daft, but it reveals how politicians are searching for familiar frameworks to explain unfamiliar technologies.

Economic Displacement

Waters argues that AI will only become a dominant political issue when its economic effects become undeniable. Currently, the concentration of wealth and power in Silicon Valley is stark, but the impact on ordinary workers is still nascent. “At some point, AI could really cut into jobs or depress earnings,” he writes.

The political tipping point may arrive when high-status professions begin to feel the squeeze. If large numbers of lawyers, consultants, or analysts find their roles automated, the backlash could be swift and severe. “Washington is full of lawyers,” Waters notes. “It wouldn’t feel as remote as coal miners in West Virginia.”

For now, however, the existential dread of AI remains a niche concern. The midterms will likely be fought on traditional grounds, with AI serving as a backdrop rather than a central theme. But as the technology evolves and its economic disruptions deepen, the phantom menace may well become a very real political force in the next election cycle. Until then, politicians will continue to grapple with how to regulate a technology they barely understand, while voters wait to see if the robots will take their jobs or their lives.