There is an old saying that you can learn a lot about people by watching what they refuse to believe. In 2026, the American electorate has turned that idea into a national pastime. We now measure our civic health not by whether elections are legitimate, but by whether the outcome happens to match our vote. The familiar cliché of a functioning democracy is that citizens trust the process even when they lose. But a new survey of more than 2,000 American adults suggests that in the run-up to the 2026 midterms, trust is no longer a baseline civic assumption. It has become a winner’s privilege. The data shows an electorate that is exhausted, suspicious, and increasingly willing to outsource its judgment to machines, while simultaneously being fooled by them. We are not living in a moment of shared facts. We are living in a moment where facts are weapons, where the same election can be seen as a triumph by one side and a theft by the other, and where both sides are convinced the other is manufacturing the lies. The real story of this election cycle is not just polling averages or congressional maps. It is the quiet collapse of confidence in the very machinery that is supposed to turn votes into voices.
The most striking finding is how conditional trust has become. When Democrats are asked whether they would trust a Democratic House victory, 92.5% say they would trust it completely or mostly. But if a Republican wins the House, that number drops to a shaky 34.3%. Republicans display the same elasticity, though slightly less extreme: 83.5% would trust a Republican win, while only 45.9% would trust a Democratic one. Independents, the swing voters who often decide these elections, land somewhere in the middle but are still full of doubt: 70.3% trust a Democratic victory, while 44.7% trust a Republican one. These numbers are not just dry statistical artifacts. They are a portrait of a citizenry that has stopped evaluating the integrity of elections and started evaluating the desirability of the result. More than three in five Democrats report low or no trust in a Republican-controlled House, and half of Republicans feel the same way about a Democratic one. None of this means these voters will refuse to accept the outcome on Inauguration Day. But it means the reservoir of goodwill that democracies depend on is dangerously shallow. Institutional confidence is no longer standard issue. It is a reward for holding the winning ticket. In other words, we don’t trust elections because they are fair. We trust them because we won.
When it comes to misinformation, the electorate is certain of one thing: everyone is lying, but the other side is lying more. Across the full sample, 57.2% of respondents blame Republicans for spreading more campaign misinformation, while 17.2% blame Democrats and 20.8% blame both equally. Independents, the group that actually swings midterms, show their hand clearly: nearly half of them, 49.6%, point their finger at the GOP, while only 11.8% blame Democrats. But there is a painful gap between identifying a liar and being willing to punish one at the ballot box. When asked if they would still consider voting for a candidate who deliberately spread false information, the results are telling. Republicans are the most forgiving, with 17.1% saying they would still back such a candidate. Democrats are more restrained, with 9.4% saying the same. Independents turn out to be the harshest judges, with just 5.7% willing to look past deliberate deception. What this suggests is that dishonesty has become so expected in politics that for many voters, it is not a disqualifier but a feature. We have moved from wanting honest leaders to merely wanting dishonest leaders on our side. The partisan brain is remarkably flexible when it comes to moral standards: a lie told by our candidate is a strategy, while a lie told by their candidate is a crime. The only people still holding the line are the independents, and even their patience is limited.
The generational story is almost counterintuitive. We have been told for years that older voters are the most vulnerable to online misinformation, that they are the ones forwarding hoaxes and falling for internet scams. But the survey data tells a very different story. When asked whether false or misleading information has made it harder to choose a candidate, the difficulty strictly decreases with age. More than half of 18-to-24-year-olds, 53.0%, say the information environment is making their vote choice harder. Among voters 65 and older, that number collapses to 27.8%. Gen Z is not navigating the digital landscape with ease. They are suffocating in it. They grew up online, but that does not mean they are immune to the noise. In fact, it means they are drowning in more of it than any generation before them. And they are not just passive consumers. A fifth of young adults, 20.4%, admit to sharing political information online that they later discovered was false, compared to just 8.2% of the oldest voters. This is not a story of youthful foolishness. It is a story of algorithmic chaos. Young people are trying to make political decisions in an environment where every tweet, every headline, every deepfake and half-truth is fighting for their attention. The older generation, with fewer social platforms and more time to develop pattern recognition, may simply be better at ignoring the noise. The young, meanwhile, are trying to hear a whisper in a hurricane.
Into this chaos has stepped a new kind of referee: artificial intelligence. More than a third of respondents, 33.8%, say they have used AI for at least one political task during the midterm cycle. They are not just asking chatbots where to vote. They are asking them what to think. About 21.5% have used AI to check whether a political claim is true, 19.5% to find candidate or policy information, 12.1% to compare candidates or parties, and 3.5% to ask for a recommendation on who to support. Men are relying on this technology more than women, with 39.7% using AI for some political task compared to 28.5% of women. The irony is brutal: voters are outsourcing their search for truth to a machine at the exact moment that machine is being used to deceive them. More than a third of respondents, 36.7%, report that they have mistaken AI-generated political content for genuine material. The public knows the dangers. When asked about their concerns, 63.1% point to fabricated events and 61.2% to false policy claims, while fake celebrity endorsements barely register at 8.8%. What we are seeing is a desperate attempt to climb out of a hole by using the same tool that helped dig it. The electorate is deploying algorithms to protect itself from algorithmic deception, with all the instability that entails. It is as if voters are trusting a map that keeps redrawing itself, hoping that this time it leads somewhere solid.
If there is a unifying theme in all this data, it is that trust is not a value. It is a survival mechanism. When the economy, prices, and jobs are the top suspected areas of misinformation at 72.8%, followed by candidates’ personal lives at 59.0% and immigration at 56.2%, we see that voters are not merely confused. They are protecting what little certainty they have left. State-level numbers offer even more color: Kansas leads the nation with 81.0% suspecting immigration-related misinformation. Maryland reports the highest rate of voter confusion at 60.5%. South Carolina shows the strongest pattern of AI adoption, with 54.3% using AI for midterm information and 40.0% for fact-checking. In New Jersey, 48.0% have mistaken AI-generated content for the real thing. In Minnesota, voters are more worried about candidate character than economic claims. These figures are not just charts and percentages. They are the sound of a country trying to find its footing in a hall of mirrors. The 2026 midterms are not just an argument over policy or governance. They are an argument over who gets to define reality. And the real crisis is not that Americans cannot agree on the answer. It is that we have stopped believing any answer can be trusted unless it comes from our own side. We want to believe in a system that believes in us. But for now, we are simply hoping that the machine we built to save us is not the same machine that buries us.