In the final stretch before the Nov. 3 midterm elections, millions of Americans are doing something their parents and grandparents did before them: trying to figure out what is true, who to trust, and how to cast a ballot that reflects their values. But this year, the road to the voting booth feels different. It is paved not just with campaign ads and yard signs, but with algorithmically amplified outrage, social media posts that look real but are not, and a low-grade anxiety about what might happen if you say the wrong thing online. Democracy has always been messy, but the digital age has made it uniquely confusing. Against that backdrop, two researchers at the University of Colorado Boulder are asking a deeply human question: how is technology shaping whether people feel willing and able to participate in civic life? Jason Thatcher, a professor of information systems at CU Boulder’s Leeds School of Business, studies how artificial intelligence, misinformation, and online harassment influence political participation. Susan Abbott, an assistant teaching professor of media studies, examines how those same forces affect civic engagement and what can be done to strengthen democracy. Together, their work paints a picture of an election season in which the real battle is not just between candidates, but between clarity and confusion, participation and withdrawal.
Jason Thatcher’s research begins with a simple observation: the way people get information about politics has been completely transformed in a single generation. It used to be that voters relied on local newspapers, televised debates, and conversations with neighbors. Today, many people’s political reality is constructed by algorithms that decide which posts to show, which ads to serve, and which stories to suppress. That can be empowering, because it allows diverse voices to be heard. But it also means people are more vulnerable to falsehoods, manipulated images, and highly targeted disinformation campaigns. Thatcher has focused specifically on how voters respond to AI fact-checking, and his research has yielded a sobering conclusion: AI fact-checking works, but mostly for progressives. In other words, when a computer-generated correction labels a claim as false, some people update their beliefs, while others dig in deeper. This is not because one group is smarter than another. It has to do with trust, identity, and the psychological discomfort of having your worldview challenged. If a person sees a fact-check from a source they already distrust, or if the correction conflicts with their political identity, the brain often protects the original belief rather than accepting the new information. The result is a paradox: a tool designed to reduce misinformation can actually widen the partisan divide, because it is experienced differently by different people.
Why does misinformation persist even after it has been fact-checked and corrected? This is one of the most frustrating questions in modern democracy, and Thatcher’s work offers important explanations. Psychological research has long shown that people do not simply delete false information from their minds when presented with the truth. Instead, the original false claim often remains memorable and influential, especially if it has been repeated many times. This is sometimes called the continued influence effect. A false statement can keep shaping a person’s attitudes even after they consciously accept a correction. Social media makes this worse by making misinformation feel familiar and widely accepted. When a lie is shared by friends, liked by strangers, and repeated across platforms, it begins to feel like common knowledge. The more often people encounter a claim, the more true it seems, regardless of the actual evidence. That is why a lie that is repeated a thousand times can seem more believable than a correction that appears only once in a small font under a post. For ordinary voters, this means that simply labeling something as false is not enough. Corrections need to be emotionally resonant, repeated, and delivered by trusted voices. Otherwise, the misinformation sticks while the correction fades. Thatcher’s research underscores how important it is to understand misinformation not just as a technical problem, but as a human one, rooted in cognition, emotion, and social belonging.
Another disturbing finding from Thatcher’s research is the way political doxing can push people out of public life entirely. Doxing, the practice of publishing someone’s private information online with the intent to harm, has become a common weapon in our polarized political environment. It might begin with a comment at a school board meeting or a petition you sign to support a local library. Then, someone with a grievance finds your home address, phone number, or employer and posts it for the world to see. The result is not just embarrassment; it is fear. People stop attending municipal meetings, stop posting opinions, and stop even telling their neighbors how they plan to vote. Thatcher’s research shows that the fear of doxing is not limited to public figures or political activists. Ordinary citizens, especially women, people of color, and those in vulnerable communities, are changing their behavior to avoid becoming targets. This is a quiet crisis for democracy. When people withdraw from public spaces, the public conversation becomes smaller, less diverse, and less representative. A healthy democracy depends on participation, not just on election day, but in the everyday work of civic life: attending forums, volunteering for campaigns, writing letters to the editor, and speaking up at city hall. When people are afraid to do those things, the system loses the very voices that make it legitimate. It is not enough to protect voters from foreign interference or hacked voting machines; we also have to protect them from the chilling effect of harassment by their own neighbors.
Susan Abbott, the media studies scholar, approaches these problems from a different but complementary angle. She looks at the ways AI-generated content and social media misinformation shape whether people feel they have a stake in civic life. For many people, the constant exposure to outrage, lies, and negativity online creates a sense of hopelessness. They feel that the system is rigged, that their participation will not matter, and that they are surrounded by enemies rather than fellow citizens. Abbott argues that one of the most important solutions is media literacy, but not the kind that simply tells people to fact-check before sharing. True media literacy is more human than that. It is about teaching people to ask why a piece of information was created, who benefits from them believing it, and what evidence supports it. It is about giving people the confidence to think critically without becoming cynical, and helping them recognize that their own emotional reactions are being engineered by algorithms designed to keep them watching, clicking, and angry. But Abbott also sees promise in structural approaches to healing democratic culture, especially citizen assemblies and deliberative polling. These are not abstract theories; they are practical ways of bringing ordinary people together to learn about issues, hear from a range of experts, and then talk through their differences in a structured, respectful environment. Unlike social media, where the loudest and most extreme voices dominate, deliberative processes give everyone a chance to be heard. They remind participants that people who disagree with them are not monsters, but human beings with different experiences and concerns. This kind of engagement does not make partisan divisions disappear, but it does make them manageable and less dehumanizing.
As the midterms approach, the research from Thatcher and Abbott offers both a warning and a reason for hope. The warning is clear: artificial intelligence, misinformation, and digital harassment are not side issues in American democracy. They are shaping who votes, who speaks, who stays silent, and who decides to walk away from public life altogether. If we do not take these forces seriously, we risk building a democracy that looks free on the outside but is hollow on the inside, because too many people are too frightened, exhausted, or confused to participate. The reason for hope is that these problems are not inevitable. Human beings have the capacity to adapt, to learn, and to create institutions that encourage the better angels of our nature. We can build fact-checking tools that are more trusted across the political spectrum. We can develop correction strategies that account for the way our minds actually work. We can create laws and cultural norms that make doxing socially unacceptable and legally punishable. We can teach media literacy in schools, workplaces, and community centers. And we can experiment with citizen assemblies and deliberative polling to give people a meaningful way to engage with one another beyond the platform-driven outrage machine. In the end, the midterm elections are not just about policies or candidates. They are about whether ordinary Americans still believe that their voice matters, that their participation is safe, and that the democratic process is worth their time. The answer depends on all of us, not just the experts, and not just on election day. It depends on how we treat one another online and offline, how we handle uncomfortable information, and whether we are willing to defend not only our own right to speak, but the right of people we disagree with to speak as well. That is the real work of citizenship, and it is work that will continue long after the votes are counted.