In the modern political world, truth faces a two-front war: artificial intelligence can fabricate lies, and it also gives real liars an excuse to deny the truth. From robocalls imitating President Biden’s voice to deepfake videos of foreign leaders, AI is already impacting elections and international politics. At the same time, politicians can now dodge scandals with a degree of plausible deniability by blaming AI. Together, actual AI disinformation and using AI as a scapegoat degrade two foundational traits necessary for a functioning democracy – access to truth and accountability.
AI’s threat to democracy lies not only in the fictions it produces, but also in the way it lets liars dismiss truth as counterfeit. Legal scholars Robert Chesney and Danielle Citron warned back in 2019 that deepfake technology wouldn’t just spread disinformation – it would make it easier for liars to avoid accountability. They called this the “liars dividend.” The concept is simple: once people know convincing fakes are possible, a politician caught in a real scandal can simply wave it off as fabricated. Last week, a video went viral of something being tossed out of an upper story White House window. As reported by the Independent, President Trump blamed AI, despite its veracity being confirmed by his team several hours prior. Similarly, the very same day in Caracas, Venezuelan Communications Minister Freddy Ñáñez dismissed video evidence of a US military strike on a Venezuelan gang’s vessel, describing the video as “almost cartoonish animation.” Reuters, however, has so far found no evidence of manipulation. These parallel incidents highlight how leaders – across different nations, holding different positions – are using AI as a convenient scapegoat for both trivial and serious matters. This is the danger of the liar’s dividend: it makes any disbelief plausible.
At the same time, AI isn’t just an excuse — it is a weapon of disinformation, spreading lies faster than they can be debunked (and casting even fact-checks into doubt.) According to Wired, the amount of content pumped out by Russian disinformation campaigns has skyrocketed since Sept. 2024, likely driven by consumer-grade AI tools. Organized disinformation has become more scalable and efficient, increasing potential national security concerns regarding engineered polarization and election interference. Several groups linked with the Russian government, like CopyCop, have even used LLMs to make clones of legitimate media outlets designed to disinform. Intraparty politics have even begun using AI to cheat.
As reported by NBC News, Steven Kramer, a political consultant working for Rep. Dean Phillips, D-Minn, commissioned a robocall that impersonated President Biden using AI during the 2024 New Hampshire Democratic Primary. The call, which reached over 5000 New Hampshire voters, used President Biden’s voice to tell voters not to waste their vote in the primaries. As a challenger backed by a smaller but more passionate base, Rep. Phillips stood to benefit if overall turnout dropped, since lower participation would hurt Biden’s broad support more than his own base. Once the robocalls were noticed by the authorities, Rep. Phillips quickly disavowed any knowledge or connection to them, condemned the act, and severed ties with Kramer. Since then, Kramer has claimed full responsibility for the act, which he claims was both independent from his client and was an “act of civil disobedience to call attention to the dangers of AI in politics.” Around the same time the robocalls went out, Phillip’s campaign had paid Kramer over $250,000, which raises doubts on his claim that it was merely a public warning. Whatever investigators ultimately conclude, the story underscores a larger truth — AI deception is already reshaping the political playing field.
The problem with AI isn’t simply fake content or convenient scapegoats; together, they undermine the very foundations of democratic governance. Among the core principles identified by political theorists, two are especially vital here: citizens must have access to reliable information so they might make informed choices, and leaders must be held accountable for their actions as Dahl argues in On Democracy. Without accountability, leaders cease to represent their constituents and pursue only their own interests. And if any evidence of wrongdoing or misrepresenting the will of their constituents can be brushed away as fake, voters lose the ability to distinguish who truly represents them when election time comes.
With AI, both of these principles are at risk. Disinformation campaigns, whether foreign or domestic, flood American media, creating convincing falsehoods. As the line between truth and fiction blurs, people gravitate toward the narratives that confirm their opinions, eroding trust even in genuine facts. The quality of AI is unlikely to ever get worse — as time goes on, it will only become more difficult to distinguish AI from real life. With this disappearing line, politicians will increasingly find it easy to dismiss compromising photos or incriminating evidence as AI. The result is a steady erosion of democratic debate, electoral integrity and public trust. With a loss of shared reality, polarization will harden, leaving democratic systems vulnerable to cynicism and instability.
Any serious response walks a fine line: protect the public from malicious use of AI without giving the government the power to dictate truth itself. A cure that undermines free expression and democratic trust could be worse than the disease. Governments, media, and citizens must respond decisively if democracy is to withstand this threat, but that response must prioritize transparency and accountability rather than censorship or control. The challenge is immense, but failing to act, or acting wrongly, would mean accepting the slow death of accountability itself.
