AI-generated misinformation is eroding democracy worldwide

AI-generated misinformation is eroding democracy worldwide

Artificial intelligence is no longer just a future concern for democracies — it is a present danger. In at least 50 countries, AI-generated content has already disrupted elections, misled voters, and undermined faith in democratic institutions. From fake videos and voice clones to manipulated news reports, the spread of synthetic media has reached a tipping point, reshaping how elections are fought and won, the New York Times reported.

From deepfakes to disinformation

In recent elections across Canada, Poland, Romania, Argentina, South Korea, and India, AI was used to produce deceptive images, videos, and social media posts targeting candidates. In Canada, an AI-generated photo falsely showed candidate Mark Carney with disgraced financier Jeffrey Epstein. In Poland, a fake video of Donald Trump endorsing a far-right leader circulated widely. And in Romania, a wave of AI-based disinformation prompted courts to annul the first round of voting — the first known case of AI leading to a cancelled election.

According to the International Panel on the Information Environment, AI was used in over 80% of the world’s elections in 2024. Of the 215 cases documented, nearly 70% had harmful consequences. This year, AI manipulation has continued in at least nine major elections, from Australia to Moldova.

Foreign interference made easier

Autocratic states like Russia, China, and Iran are exploiting generative AI tools to meddle in foreign elections more effectively than ever before. In Moldova, AI-generated videos praising pro-EU politicians were traced back to Russian influence campaigns. In Germany, researchers found a bot account using OpenAI’s tools to support the far-right Alternative for Germany party, which doubled its parliamentary seats.

These operations no longer need troll farms or clumsy propaganda. With tools like Midjourney, Veo, and Grok, high-quality disinformation can be produced at scale and tailored to local contexts, making it harder for platforms and voters to distinguish truth from fabrication.

The double edge of campaign tech

Not all uses of AI in politics are malicious. In about 25% of cases, candidates have used AI to translate speeches, target voters, or even clone themselves to motivate campaign workers. In India’s recent election, candidate avatars became commonplace.

Yet this same accessibility has made harmful uses rampant. In the U.S., AI-manipulated videos during the 2024 presidential election drew warnings from multiple intelligence agencies. But under President Trump’s administration, key teams at the FBI and Cybersecurity and Infrastructure Security Agency working on election integrity have been disbanded, weakening defences.

Tech platforms struggle to respond

Social media companies, while having rules against AI abuse, have struggled to enforce them. A study by the University of Texas at Austin found that Meta failed to properly label most AI-generated content during India’s election. Meanwhile, a University of Notre Dame study showed that AI-created accounts easily bypass detection on platforms like Facebook, Reddit, TikTok, and X.

In Romania’s recent election, TikTok became a hub for manipulated content, including fake videos of candidates and cloned voices of international figures like Trump. TikTok says it removed over 7,000 posts in the lead-up to the second round of voting, but the European Union has launched an investigation into whether the platform did enough.

Experts warn of long-term damage

Lucas Hansen of the nonprofit CivAI says the problem goes beyond deception. The flood of synthetic content is polluting the information ecosystem so thoroughly that citizens may stop trusting any political information. “The pollution of the information ecosystem is going to be one of the most difficult things to overcome,” he said.

With tools now able to generate high-quality disinformation quickly and anonymously, democracies are facing an existential test. As Professor Madalina Botan of Romania put it, “What can compete with this?”

The answer, experts warn, may determine whether democratic societies can survive the next wave of global elections.