As Americans head to the polls for the midterm elections, a quiet but corrosive force is shaping the way people read, share, and ultimately trust political information online. It isn’t just misinformation, foreign interference, or partisan vitriol—though all of those play a part. It’s the growing suspicion that the person behind a post, comment, or reply might not be a person at all. New research from the University of Sydney has found that accusations of artificial intelligence use have become a common rhetorical weapon in online political debates, used almost reflexively to dismiss views people disagree with. The effect, researchers warn, is an environment of deep distrust that could change how citizens engage with elections not just in the United States, but around the world. The study, led by Dr. Timothy Koskie of the University’s Centre for AI, Trust and Governance, analyzed more than 220,000 posts and comments from Reddit’s five largest news and political communities in the lead-up to the 2024 U.S. presidential election. What they found was a pattern both striking and unsettling: rather than simply debating policy or character, users frequently accused one another of being AI-generated, bots, or pawns of algorithmic manipulation. The accusation—often reduced to a dismissive “that’s AI”—has become a shorthand for saying, “I don’t have to take this seriously.” And in an age where so many people rely on online platforms to understand political issues and evaluate candidates, that reflex may be doing real damage to the foundations of democratic debate.
The power of this dynamic lies in its ability to undermine trust without offering any proof. Dr. Koskie and his team examined Reddit discussions during the turbulent months before the 2024 election, looking at how users talked about AI, bots, and algorithms in political contexts. Their findings, published in the journal Communication Research and Practice, reveal that these accusations were not random or evenly distributed. They were almost exclusively aimed at people holding contrasting political viewpoints. If you disagreed with someone, one of the easiest ways to discredit them was to suggest that their words weren’t their own—that they were generated by a machine, or planted by a coordinated campaign. This tactic, researchers note, feeds into the broader misinformation and dishonesty that have multiplied alongside social media itself. It creates a world where every post is suspect, where every argument can be dismissed as manufactured, and where genuine human opinion gets caught in the crossfire. What makes this so dangerous is that it erodes confidence not only in the information itself, but in the very possibility of trusting what we read. If everything can be accused of being AI slop, then nothing can be believed, and the entire concept of a shared public conversation begins to collapse.
Among the most common themes in the comments and posts analyzed were references to bots and algorithms. Accusations that other users were “bots” or “paid trolls” were used to silence dissent, dismiss inconvenient facts, and portray opponents as part of a shadowy, automated conspiracy. The researchers found that conversations often lost momentum the moment someone was labeled a bot. Suddenly, the human being on the other side of the screen becomes an abstraction, an unworthy participant in democratic dialogue. Dr. Koskie noted that this pattern mirrors the “fake news” rhetoric popularized by Donald Trump, which encouraged distrust of journalists and media organizations that reported critically on him. In the same way, calling someone a bot or accusing them of spreading AI-generated content delegitimizes their voice without requiring any evidence. This has a chilling effect on political engagement. People become more reluctant to share their opinions, more suspicious of the information they encounter, and more likely to retreat into echo chambers where they only hear views that confirm their own. Over time, the public square becomes a place not of debate and deliberation, but of paranoia and hostility. And while concerns about bots and algorithmic manipulation are legitimate—foreign influence campaigns and coordinated disinformation are real threats—the weaponization of these concerns risks undermining discussions that are entirely genuine, including those that could help voters make better-informed decisions.
Another troubling discovery was how easily the mere association with AI could render human-generated content untrustworthy. In the digital environment of the 2024 election, being accused of using AI was enough to taint a post, regardless of whether it was true. Users expressed widespread suspicion about deepfakes, fabricated images, and automated text, and this suspicion bled into their perceptions of ordinary political speech. A thoughtful, carefully written comment from a real person could be dismissed as “AI-generated” simply because it seemed too polished, too detailed, or too well-structured. This is a paradoxical outcome: the very tools designed to help people communicate and create are now being used as evidence that they cannot be trusted. The researchers found that concerns about AI misuse frequently included unverifiable accusations about platform algorithm manipulation and hidden agendas. Many users worried that social media algorithms were secretly promoting or suppressing content for political or financial gain, and they viewed the platforms themselves with deep suspicion. This lack of transparency, combined with the prevalence of AI-related conspiracy theories, created a fog of uncertainty around even the most basic political questions. Who wrote this? Why is this showing up in my feed? Is this whole conversation being manufactured to make me feel a certain way? When people cannot answer these questions with confidence, their ability to engage meaningfully with elections and public debate is severely compromised. Credible information struggles to gain traction, while sensational claims and cynical accusations thrive in the confusion.
If there is a way forward, it involves recognizing that the problem is not simply technological, but social and institutional. The burden of restoring trust cannot rest on individual users alone. Social media platforms have a responsibility to promote trustworthy and transparent information environments—not just by policing obvious misinformation, but by making their algorithms more open to scrutiny and by creating clear guidelines for the use of AI in political content. Policymakers and technology companies must work together to ensure that AI tools support informed democratic participation rather than undermine it. At the same time, the study highlights the importance of maintaining access to a diverse range of independent news sources and perspectives. When people have multiple reliable places to turn for information, they are less dependent on the chaotic and often manipulative dynamics of social media feeds. Dr. Koskie emphasizes that we should all be critical of the information we encounter online, but he warns that genuine concerns about AI are being weaponized to suppress legitimate democratic communication. The balance is delicate. On one hand, voters need to be vigilant about the risks of AI-driven deception. On the other, they cannot let that vigilance curdle into a reflexive cynicism that treats every human opinion as suspect. That kind of cynicism, the research suggests, is not just unhelpful—it is actively harmful to democratic debate at a time when robust, honest, and open discourse is more important than ever.
Ultimately, this study is a warning about the fragility of trust in the age of artificial intelligence. With major elections taking place around the world, from the United States to Europe to Asia and beyond, the relationship between AI, trust, and democracy is one of the defining challenges of our time. The 2024 U.S. election may have been the focus of the research, but its lessons apply far beyond Reddit or American politics. Everywhere, people are navigating a digital landscape in which the line between human and machine is blurring, and in which distrust is often easier than understanding. The findings from the University of Sydney should serve as a call to action for platforms, governments, and citizens alike. We need to build environments where credible information can thrive, where people can disagree without accusing each other of being bots, and where AI tools are used to enhance rather than erode public discourse. That will require more than technical fixes; it will require a renewed commitment to the values that make democracy work—curiosity, humility, empathy, and a willingness to listen to perspectives we don’t already share. The future of democratic debate may depend less on what AI can do, and more on how we choose to treat one another in the space between a post and a reply. In that fragile space, trust is everything. And once it is lost, it is very hard to get back.