The Confirmation Trap: How Cognitive Bias Corrodes American Politics

The Incident: A Case Study in Real Time

  <p>In September 2025, during an episode of <em>The Joe Rogan Experience</em> with comedian Tim Dillon, Joe Rogan showed what he believed was a real video of Minnesota Governor Tim Walz dancing down an escalator in a profane anti-Trump shirt. The video was AI-generated — Walz's face had been superimposed onto the body of a TikTok creator. It was, by most accounts, obviously fake.<sup><a href="#s1">[1]</a></sup></p>

  <p>When Rogan's producer Jamie Vernon told him "No, it's not real," Rogan's first response was to insist otherwise: "Yes it is." He called Vernon a "plant." The moment was significant not because a podcast host fell for a deepfake — that's increasingly common — but because of what happened next. After finally acknowledging the video was fake, Rogan offered this defense:<sup><a href="#s2">[2]</a></sup></p>

  <blockquote>
    I fell for it too, and do you know why I fell for it? Because I believe that he's capable of doing something that dumb.
    <cite>— Joe Rogan, <em>The Joe Rogan Experience</em>, September 2025</cite>
  </blockquote>

  <p>This single sentence is a near-perfect illustration of confirmation bias in action. Rogan did not evaluate the video on its technical merits. He did not check whether it was real. He believed it because it was consistent with his existing view of Walz — and when confronted with the truth, he defended his credulity by appealing to that same prior belief. The error didn't prompt reflection; it prompted rationalization.</p>

  <p>The OECD formally classified this as an AI incident, noting that Rogan's insistence on the video's authenticity — even after correction — contributed to the spread of misinformation to his massive audience.<sup><a href="#s3">[3]</a></sup></p>

  <h2>What Is Confirmation Bias?</h2>

  <p>Confirmation bias is the tendency to search for, interpret, favor, and recall information in a way that confirms one's preexisting beliefs. The term was coined by British psychologist Peter Wason in the 1960s, whose experiments demonstrated that people overwhelmingly seek evidence that supports their hypotheses rather than evidence that might disprove them.<sup><a href="#s4">[4]</a></sup></p>

  <div>
    <div>Landmark Experiment</div>
    <h4>Wason's 2-4-6 Task (1960)</h4>
    <p>Participants were told the number sequence 2, 4, 6 fit a rule. They were asked to discover the rule by proposing new sequences. The actual rule was simply "any ascending sequence" — but participants consistently tested only sequences that confirmed their more specific guesses (like "increases by 2"), rarely testing sequences that might disprove them. Most never found the correct rule.</p>
  </div>

  <p>Confirmation bias operates through three reinforcing mechanisms:<sup><a href="#s4">[4]</a></sup></p>

  <p><strong>Biased search.</strong> We look for information that confirms what we already believe. Rogan didn't fact-check the Walz video before sharing it with millions of listeners — he didn't need to, because it already fit his mental model.</p>

  <p><strong>Biased interpretation.</strong> We interpret ambiguous information in ways that support our existing views. The same economic data can look like recovery to one partisan and decline to another. A 2020 study in <em>American Politics Research</em> by Acland and Lerman found "substantial and statistically significant confirmation bias" in how partisans evaluated identical information about government services — with both Democrats and Republicans rating the same service performance differently depending on which party they believed was responsible.<sup><a href="#s5">[5]</a></sup></p>

  <p><strong>Biased memory.</strong> We remember information that supports our beliefs more readily than information that contradicts them. Over time, this creates an increasingly distorted picture of reality that feels increasingly certain.</p>

  <h2>When Belief Becomes Identity</h2>

  <p>In American politics, confirmation bias doesn't just shape how people process information — it is amplified by something more powerful: identity.</p>

  <p>Yale law professor Dan Kahan's research on "identity-protective cognition" shows that people don't just prefer confirming information — they are driven to defend beliefs that are central to their group identity. This goes beyond simple bias. People will unconsciously dismiss evidence that conflicts with their group's positions, even when they have the cognitive skills to evaluate that evidence correctly.<sup><a href="#s6">[6]</a></sup></p>

  <p>Kahan's most striking finding involves numeracy: in his experiments, people who scored high on mathematical reasoning were <em>more</em> likely to get politically charged math problems wrong when the correct answer contradicted their partisan beliefs. High-numeracy partisans selected answers that affirmed their political outlooks even when those answers were mathematically incorrect. Cognitive sophistication didn't reduce bias — it enhanced it.<sup><a href="#s6">[6]</a></sup></p>

  <div>
    <div>Key Finding</div>
    <h4>Smarter ≠ Less Biased</h4>
    <p>Kahan's research demonstrates that intelligence and education do not inoculate against confirmation bias. In politically charged contexts, greater cognitive ability can actually make people <em>better</em> at constructing sophisticated rationalizations for their pre-existing beliefs — not better at finding the truth.</p>
  </div>

  <p>This has profound implications. It means the problem isn't simply that some people are uninformed. Both conservatives and liberals demonstrated the same tendency to fit empirical evidence to their ideological predispositions. People are also more likely to accept misinformation and resist correction when that misinformation is identity-affirming rather than identity-threatening.<sup><a href="#s6">[6]</a></sup></p>

  <h2>The Fragile Ego Problem</h2>

  <p>Rogan's response to being fact-checked — doubling down, then rationalizing — is not unusual. But the psychology behind it reveals something deeper than stubbornness.</p>

  <p>In a 2018 article in <em>Psychology Today</em>, psychologist Guy Winch explained why some people cannot admit they are wrong. His answer: a fragile ego masquerading as strength.<sup><a href="#s7">[7]</a></sup></p>

  <blockquote>
    Some people have such a fragile ego, such brittle self-esteem, such a weak "psychological constitution," that admitting they made a mistake or that they were wrong is fundamentally too threatening for their egos to tolerate. Accepting they were wrong, absorbing that reality, would be so psychologically shattering that their defense mechanisms do something remarkable to avoid doing so — they literally distort their perception of reality to make it less threatening.
    <cite>— Guy Winch, Ph.D., Psychology Today, November 2018</cite>
  </blockquote>

  <p>Winch makes a critical distinction that is often misunderstood: what looks like confidence is actually compulsion. "To the outside world, they look as if they're confidently standing their ground and not backing down, things we associate with strength," he writes. "But psychological rigidity is not a sign of strength, it is an indication of weakness. These people are not <em>choosing</em> to stand their ground; they're <em>compelled</em> to do so in order to protect their fragile egos."<sup><a href="#s7">[7]</a></sup></p>

  <p>This distinction matters enormously in American political culture, which prizes conviction and decisiveness. When public figures demonstrate an inability to admit error, it is frequently interpreted as strength of character rather than what the research suggests it actually is: a defense mechanism against psychological fragility.</p>

  <p>Consider the Rogan exchange in this light. When confronted with the fact that the video was fake, Rogan did not say, "I was wrong, I should have checked." Instead, he defended his credulity by pointing to the very belief that caused the error — that Walz was "capable of doing something that dumb." The error didn't update his beliefs; it reinforced them. That is not confidence. It is a closed loop.</p>

  <h2>Scientific Thinking vs. Confirmation Seeking</h2>

  <p>There is a fundamental difference between two modes of engaging with information, and this difference is at the heart of the crisis in American political reasoning:</p>

  <div>
    <div>
      <h4>Scientific Thinking</h4>
      <ul>
        <li>Form a hypothesis, then try to <em>disprove</em> it</li>
        <li>Seek disconfirming evidence actively</li>
        <li>Update beliefs when evidence demands it</li>
        <li>Treat being wrong as information, not failure</li>
        <li>Require replication and peer review</li>
        <li>Provisional conclusions, always revisable</li>
      </ul>
    </div>
    <div>
      <h4>Confirmation Seeking</h4>
      <ul>
        <li>Start with a conclusion, then find support</li>
        <li>Avoid or dismiss contradicting evidence</li>
        <li>Treat challenges to beliefs as personal attacks</li>
        <li>Equate being wrong with being weak</li>
        <li>Accept anecdotal evidence that "feels right"</li>
        <li>Fixed conclusions, defended at all costs</li>
      </ul>
    </div>
  </div>

  <p>The philosopher Karl Popper argued that the defining feature of scientific knowledge is not that it can be proven true, but that it can be proven false. For a theory to be scientific, it must make predictions that could, in principle, be disproven. Each falsification is not a failure — it is progress, telling us definitively that a particular explanation is wrong and pushing inquiry toward better ones.<sup><a href="#s8">[8]</a></sup></p>

  <p>Popper emphasized that "knowledge progresses through guesses, tentative solutions, and conjectures, which are then controlled by criticism and attempted refutations." The critical spirit — the willingness to subject your own beliefs to rigorous testing — is, in Popper's view, the very essence of rationality.<sup><a href="#s8">[8]</a></sup></p>

  <p>This is the exact opposite of what confirmation bias produces. A person in the grip of confirmation bias doesn't seek refutation — they seek validation. They don't ask "could I be wrong?" but "who agrees with me?" The difference between understanding the scientific research process and simply finding agreeable facts is the difference between knowledge and the <em>feeling</em> of knowledge.</p>

  <p>Research on the Dunning-Kruger effect reinforces this point. People who overestimate their own knowledge are more susceptible to believing false news, more likely to discount expert-backed information, and more resistant to factual corrections — even when those corrections present clear counter-evidence. Social media amplifies this dynamic: social media use for news is positively related to knowledge miscalibration, the mismatch between how much people think they know and how much they actually know.<sup><a href="#s9">[9]</a></sup></p>

  <p>The problem is not a lack of available information. It is a failure to engage with information the way a scientist would — provisionally, critically, with the expectation of being wrong and the willingness to update.</p>

  <h2>How AI Exploits the Confirmation Trap</h2>

  <p>Artificial intelligence has weaponized confirmation bias. The technology to create convincing fake videos, images, and audio is now cheap, fast, and widely accessible. By 2024, AI tools could generate hyper-realistic content within seconds at minimal cost.<sup><a href="#s10">[10]</a></sup></p>

  <p>But the real danger isn't the technology itself — it's the cognitive vulnerability it exploits. AI-generated disinformation is effective not because it's indistinguishable from reality, but because people <em>want</em> to believe it. The Walz deepfake was, by many accounts, obvious. One social media commenter described it as "low-grade slop made with a Canva knockoff."<sup><a href="#s2">[2]</a></sup> It didn't need to be convincing. It only needed to confirm what the viewer already believed.</p>

  <p>This dynamic is backed by research. A 2025 report by CIVICUS and DDI Research on deepfakes in elections found that AI-generated political content shapes citizen perceptions through the mechanism of preexisting beliefs — people judge the plausibility of synthetic content based on whether it aligns with their existing political attitudes, not on its technical quality.<sup><a href="#s10">[10]</a></sup></p>

  <div>
    <div>The Paradox of AI Disinformation</div>
    <h4>Low Quality, High Impact</h4>
    <p>The most dangerous AI-generated political content may not be the most sophisticated. Crude deepfakes that confirm widely-held partisan assumptions can spread faster and be believed more readily than technically impressive fakes about implausible scenarios. Confirmation bias is the delivery mechanism; AI is merely the payload.</p>
  </div>

  <p>Research by the Knight First Amendment Institute at Columbia University, analyzing 78 election-related deepfakes, found that political misinformation fundamentally is not an AI problem — it is a human cognition problem that AI tools happen to make cheaper and faster to exploit.<sup><a href="#s11">[11]</a></sup></p>

  <p>The World Economic Forum's Global Risks Reports for 2024 and 2025 identified AI-driven disinformation as one of the most significant near-term threats to society.<sup><a href="#s10">[10]</a></sup> But the reports' framing misses something important: the threat is not that AI creates false content, but that humans are psychologically primed to consume it uncritically.</p>

  <h2>The Amplification Problem</h2>

  <p>When someone with a large audience falls for misinformation and then rationalizes it rather than retracting it, the damage compounds.</p>

  <div>
    <div>
      <span>50M+</span>
      <span>Combined followers across platforms</span>
    </div>
    <div>
      <span>14.5M</span>
      <span>Spotify followers alone</span>
    </div>
    <div>
      <span>21.9M</span>
      <span>Unique US viewers (Oct 2024)</span>
    </div>
  </div>

  <p>Joe Rogan's podcast is the most popular on Spotify, with over 14.5 million followers. His combined audience across YouTube, Spotify, and Instagram exceeds 50 million. In October 2024, he was the 34th largest media creator by unique U.S. viewers across YouTube and Facebook, reaching 21.9 million people — his highest total in 13 months. The show's audience skews male (80%) and young (56% aged 18-34).<sup><a href="#s12">[12]</a></sup></p>

  <p>This matters because when an influential figure models confirmation bias — believing something without verification because it aligns with his prior views, then defending that credulity instead of correcting it — he is not just sharing misinformation. He is modeling a way of thinking. He is teaching millions of viewers that it's acceptable to believe things without evidence, and that when you're caught being wrong, the correct response is rationalization rather than revision.</p>

  <p>This is not an isolated event. Rogan has a documented pattern. During a previous episode, he misattributed a Trump gaffe to Biden — and was similarly fact-checked by his producer on air. He has hosted climate contrarians presenting long-debunked claims as legitimate scientific debate, providing what NPR described as "a sense of false balance, like there's two sides to the scientific evidence when there is not."<sup><a href="#s13">[13]</a></sup></p>

  <p>West Virginia Representative Riley Moore also reposted the same fake Walz video before being corrected. The pattern repeats because the psychology permits it: when your audience shares your priors, there is no social cost to being wrong in the "right" direction.<sup><a href="#s2">[2]</a></sup></p>

  <h2>The Case for Intellectual Humility</h2>

  <p>If confirmation bias is the disease, the research consistently points to one antidote: intellectual humility — the recognition that one's knowledge is limited and that one might be wrong.</p>

  <p>This is not weakness. The empirical evidence shows the opposite.</p>

  <p>Research published in the journal <em>Personality and Individual Differences</em> found that intellectual humility is rooted in cognitive flexibility — the ability to adapt thinking in response to new information. High cognitive flexibility predicted intellectual humility, and both traits were associated with stronger reasoning ability.<sup><a href="#s14">[14]</a></sup></p>

  <p>Researchers at UC Davis and the University of Pittsburgh found that intellectually humble individuals exerted greater effort when initially failing to master material, resulting in improved memory and greater self-awareness about the limits of their own knowledge. The Decision Lab's review of this research emphasizes that intellectually humble people spend more time engaging with opposing viewpoints and are measurably less susceptible to confirmation bias.<sup><a href="#s14">[14]</a></sup></p>

  <p>Most critically for the misinformation crisis: people who score high in intellectual humility are more likely to fact-check claims before sharing them and to seek out alternative perspectives. They are measurably less susceptible to fake news — not because they are smarter, but because they are more willing to consider the possibility that what they're seeing might not be true.<sup><a href="#s14">[14]</a></sup></p>

  <blockquote>
    Psychological rigidity is not a sign of strength, it is an indication of weakness.
    <cite>— Guy Winch, Ph.D., Psychology Today</cite>
  </blockquote>

  <p>The contrast between intellectual humility and psychological rigidity maps cleanly onto the scientific mindset versus confirmation seeking. A scientist who cannot admit her hypothesis was wrong is not a good scientist — she is a bad one. A political commentator who cannot admit he was fooled is not demonstrating conviction — he is demonstrating the same fragile-ego defense mechanism Winch describes.</p>

  <p>The path forward is not about being less partisan or caring less about political outcomes. It is about developing the cognitive habits that allow people to hold strong views while remaining open to evidence that might challenge them. It is about understanding that the willingness to say "I was wrong" is not a concession — it is the single most reliable indicator of genuine intellectual strength.</p>

  <p>In Wason's 2-4-6 experiment, the participants who found the correct rule were the ones who tested sequences they believed would <em>break</em> their hypothesis. In politics, in media consumption, in everyday reasoning, the question is the same: are you looking for evidence that confirms what you already believe, or are you brave enough to look for evidence that might prove you wrong?</p>