Playbook
Tasteful Skills
“Tasteful Skills” argues that the best agent skills are not documentation or best-practice lists.
Back to Podcast Digest
AI Is Becoming America's Most Dangerous Political Issue
TL;DR
AI is turning into a real political wedge issue — host Mike Kaput says new polling from Blue Rose Research shows AI rose faster in issue importance than any topic they track, with 79% of voters worried the government has no plan for AI job losses and 77% worried entire industries could disappear.
The hosts think much of the polling is politically useful precisely because people don’t understand AI well — Paul Roetzer argues that ignorance can be weaponized, since campaigns can define AI for voters as “job loss” or “data centers ruining communities” and then use that fear to move votes.
The reassuring message from government and tech is falling flat — according to David Shor’s polling, claims like “AI won’t cause widespread job losses” have net trust of -41, and “AI will create productivity that benefits everyone” lands at -20.
Two rival AI manifestos are now openly battling for the narrative — the Pro-Human AI Declaration, backed by figures as different as Steve Bannon, Susan Rice, Richard Branson, Ralph Nader, and Yoshua Bengio, calls for halting superintelligence until it’s proven safe, while Build American AI responds: “We cannot afford to pause AI.”
Trump’s AI framework clearly favors acceleration over regulation — its seven-pillar legislative outline rejects new federal AI regulators, pushes courts rather than Congress on copyright, suggests preempting restrictive state laws, and frames access to computing power as an “American right to compute.”
The core advice from the episode is: trust no single camp on AI politics — Roetzer repeatedly tells listeners to check who funded the polling, what each group wants, and avoid defaulting to partisan silos because “neither political party knows the answer here.”
The Breakdown
Polling says AI anxiety is rising fast — but context matters
Mike opens with the headline-grabbing numbers: 79% of voters worry the government has no plan for AI job losses, 77% fear entire industries could be wiped out, and 56% worry about losing their own job. He frames it against a broader economic mood — 61% say life got less affordable in the last year, only 25% feel confident in their financial future, and just 34% think their job is secure.
Paul’s first instinct: who asked the question, and why?
Paul immediately slows the conversation down and asks the question he always asks with viral polling: who produced it, and what’s the agenda? He notes David Shor runs data science at Blue Rose Research and bluntly describes himself as someone trying to elect Democrats, which doesn’t invalidate the research — it just tells you what the research is for.
AI is “rising fast,” but it’s still not a top voter issue
Paul adds the missing denominator: AI may be the fastest-rising issue, but it still ranks just 29th out of 39 issues in this survey. The real top five remain the usual suspects — cost of living, the economy, political corruption, inflation, and healthcare — so his read is that Americans don’t actually care about AI enough yet for it to dominate elections, even if the trend line is moving quickly.
Which poll questions feel real, and which feel loaded?
He goes question by question and basically grades the polling in public. Concern that the government has no AI worker-protection plan? “100% true.” Concern that young people will face fewer job opportunities because of AI? Also real. But “entire industries being eliminated” he calls absurd, and the wage-suppression question he shrugs off as something people would fear no matter what variable you insert.
Another poll shows Americans are still barely using AI
Paul then brings in Data for Progress, a progressive think tank, and this survey is revealing in a different way: most people still aren’t deeply engaged with AI tools. He points to numbers showing broad partisan symmetry on whether people have embraced or resisted tools like ChatGPT, and laughs at the finding that 50% think they can spot AI-generated photos or videos — “they can’t.”
The AI fight now has two manifestos and some very strange alliances
Mike lays out the “dueling declarations” moment: the Pro-Human AI Declaration unites unlikely names like Steve Bannon, Susan Rice, Richard Branson, Ralph Nader, and Yoshua Bengio behind a call to prohibit superintelligence development until it’s proven safe. The counterattack, Build American AI, says a pause would backfire, weaken U.S. competitiveness, and hand strategic advantage to hostile actors.
Trump’s framework signals: build first, regulate later
The administration’s seven-pillar AI framework, they say, is short but revealing. It opposes new federal AI agencies, punts copyright to the courts, wants Congress to preempt burdensome state rules, shifts more child-safety responsibility to parents, and wraps the whole thing in “American AI dominance” — which Paul says is the pillar all the others really sit beneath.
“Choose your fighter”: abundance vs. fear
By the end, Paul says the two political messages are getting clearer: one side says AI will bring abundance and help America beat China; the other says it threatens jobs, communities, and social stability. His warning is the memorable part — don’t get trapped in your usual political silo, because campaigns will use whatever fears or fantasies keep them in power, and AI is becoming fertile ground for exactly that.
Share
Keep Reading
Two Ways to Follow
The Weekly Echo. The inbox-shaped summary of what mattered.
New editorials announced here.
