AI companies like Unanimous.ai and Expert.ai say they correctly predicted election results in battleground states while traditional polling again falls short
As traditional political polling methods appear to fall short again, some tech firms say that AI holds promise
Context & Ripple Effects
Four years after the postmortem on large-scale data analysis missing the 2016 outcome, Unanimous.ai and Expert.ai are claiming the opposite result: correct battleground-state calls while traditional polling missed again. The claim matters because election forecasting was one of the first places AI vendors could prove accuracy against a visible benchmark.
But the follow-on coverage shows the proof point didn't convert into a business: despite 30+ companies selling AI products to US campaigns, most campaigns stayed wary, per later interviews with campaigns and vendors. Meanwhile AI's role in elections migrated from prediction to participation.
First-order effects
- Unanimous.ai and Expert.ai get their strongest credibility asset — battleground-state predictions they can market against a polling industry that has now underperformed in consecutive cycles.
- Traditional pollsters face renewed pressure on methodology, with AI firms positioned as the alternative forecasting layer.
Second-order effects
- Vendors' claimed wins collide with actual demand: even after the 2020 claims, campaigns largely declined to buy AI products, so the sales pitch shifts from 'we predicted' to proving reliability to skeptical buyers.
- As trust in polls erodes further, adjacent players — voter-guides publishers and news outlets — face competition from new intermediaries, including voters using AI tools as nonpartisan researchers instead of traditional coverage.
Third-order effects
- If the pattern holds, election influence moves from forecasting claims to structural power: by the 2026 midterms, the AI industry is among the biggest financial backers of campaigns even as public anger grows over its data-center footprint.
- The longer shift is that political information — who predicts outcomes and who informs voters — migrates from pollsters and newsrooms to a small set of AI vendors whose methods are harder to audit than a published crosstab.
The trend: AI's role in US elections is expanding from contested prediction claims into campaign funding, voter-facing tools, and a direct challenge to polling and news as the default information layer.