New Polling: America's AI Reality Doesn't Add Up
Six contradictions shaping the politics of AI
We’re officially in campaign season with Texas holding its first major primaries this week, and AI is already on the trail.
Because of that, I wanted to collaborate with my friends at the Rainey Center — a public policy research organization and leadership community of elected officials across the nation — to bring you new polling on Americans' views on artificial intelligence, how people get their news, and how they feel about tech becoming more politically involved. I am also on their advisory board. This poll surveyed 1,008 registered voters via an online panel from February 25–27, 2026.*
If you want to dig deep into the numbers, here are links to the toplines and the cross-tabs. The full polling memo on the AI questions is here. Today, I want to give you my analysis of what stood out to me in this poll and a few others that have come out recently. And that is, the data consistently points to various contradictions. Here are the six that stood out to me:
The media map is fractured — and the averages lie
Newsletters are small — and disproportionately powerful
America is building the world’s AI. America is not using it.
81% want candidates with an AI plan — but distrust government AI contracts
Voters say they want neutrality — but reward perceived conviction
Data centers: education moves everyone
Before you read the breakdown, one thing worth noting: these contradictions aren’t random. They’re connected. The same fractured information environment that makes media averages misleading is the one candidates will try to navigate with AI plans they don’t fully understand, about a technology their voters are adopting unevenly, in a country that’s building the world’s most powerful AI tools while ranking 24th in actual use of them. Pull on any one of these threads, and you find the others. Here’s how I read each one.
1️⃣ The media map is fractured — and the averages lie
Why it matters: If you’re making a strategy based on topline polling numbers, you’re probably making the wrong call.
On the surface, cable and broadcast still appear to dominate as sources of political news. But once you dig into the crosstabs, the story changes.
Among voters 18–34, 81% get political news from social media. When asked where they encounter political influencers and independent commentators, YouTube leads overall at 46% — and reaches 67% of younger voters. For the first time, podcasts have surpassed talk radio in listening time (40% vs. 39%).
Meanwhile, voters 65+ answered “other” when asked where they find political influencers — at a rate of 57%. That “other” is almost certainly cable.
I started to see this clearly when I sat with the crosstabs for this poll. The overall numbers look coherent. The moment you break them by age, you’re looking at a completely different country depending on which row you’re reading.
My read: We’re no longer just in a fragmented media environment. We’re in a kaleidoscope — and it keeps turning. The number of platforms is growing. The influencers people trust are multiplying. And legacy media institutions are fighting to stay in the game at the same time. Anyone making grand statements about where Americans get their news — including me — needs to show their crosstabs.
For executives: your board may watch cable. Your current and future workforce does not. Both are important. Build a comms strategy that accounts for both realities.
2️⃣ Newsletters are small — and disproportionately powerful
Why it matters: Reach and influence are not the same thing.
Pew recently found that three in ten Americans get news from newsletters at least sometimes, with higher-income and higher-education audiences skewing toward this. In our poll, 46% of voters say they don’t read newsletters at all.
That doesn’t sound massive.
But the people who do read them tend to be journalists, Hill staff, political operatives, academics, and policy professionals. The narrative shapers.
I think about newsletters the way I used to think about Twitter. Very few Americans actually used it. But because reporters and political elites did, what happened on Twitter became the news. Substack operates the same way — smaller audience than the footprint suggests, outsized influence on what filters into mainstream coverage.
My read: Don’t use newsletters for mass persuasion. Use them for narrative formation. There’s a difference — and confusing the two is an expensive mistake.
If you've made it this far, you already know the averages are lying to you. The next four contradictions are where I show you what's underneath them — and what to actually do about it. That's what paid subscribers are here for.
