“The increasing use of deepfake technology poses a significant threat to democratic elections.”

These deepfakes are no longer crude forgeries. They are often indistinguishable from reality. In recent election cycles, AI-generated robocalls have falsely instructed voters to stay home. It is one of the fastest growing threats to election security today.

The capacity of this technology to synthesize highly convincing media systematically erodes public confidence, facilitates sophisticated political manipulation, and poses a direct, existential risk to free and fair elections. As the technology underpinning deepfakes continues its trajectory toward greater accessibility and complexity, these associated risks are projected only to escalate further.

Instances of manipulated or wholly generated content have surfaced, posing a threat to democratic discourse and electoral integrity. In September 2023, generative AI-based political interference upended Slovakia’s parliamentary elections with an audio clip allegedly featuring voices discussing election manipulation. For example, deepfakes and voice cloning have already been used to imitate candidates running for office, such as an AI-generated robocall purporting to be U.S. President Joe Biden discouraging voting.

Deepfakes can directly alter voter preferences and spread disinformation about candidates by making them appear to take policy positions they do not hold or engage in illegal behavior. Coordinated disinformation campaigns utilizing deepfake videos could prevent citizens from voting by spreading false information about election procedures or intimidating voters through blackmail.

With artificial intelligence-based methods for creating deepfakes becoming increasingly sophisticated and accessible, deepfakes are raising a set of challenging policy, technology, and legal issues. Candidates in a political campaign can be targeted by manipulated videos in which they appear to say things that could harm their chances for election.

As deepfake tools become more sophisticated and accessible, they pose a significant threat to the democratic process around the world. Policymakers must recognize the urgency of the situation and take proactive measures to address this unprecedented challenge.

Generative Artificial Intelligence (GenAI) and Large Language Models (LLMs) pose significant risks, particularly in the realm of online election interference. Deepfake technology can generate realistic videos of political figures saying or doing things they never did, eroding public trust in authentic information sources.

Deepfakes and generative AI used for disinformation undoubtedly pose a risk to fair democratic processes globally. Mitigating that risk will take a concentrated effort from stakeholders, including AI developers and operators, the press, governments, political candidates, social media platforms, and voters themselves.

Problematically, however, concern about deepfakes poses a threat of its own: unscrupulous public figures or stakeholders can use this heightened awareness to falsely claim that legitimate audio content or video footage is artificially generated and fake. The fact is, many deepfakes are poorly made and easy for humans to spot.

Deepfakes amplify misinformation's impact in unprecedented ways, erasing the line between truth and fabrication. In July 2024, a deepfake video of Kamala Harris spread rapidly, raising concerns about undetectable deceptions damaging elections. As deepfakes erode trust in content, the foundation of democracy—an informed electorate—will be at risk.

A new political ad in Georgia's U.S. Senate race is raising concerns about the use of artificial intelligence in elections after Rep. Mike Collins' campaign released a deepfake video showing Sen. Jon Ossoff mocking farmers and defending a government shutdown. Ossoff never said any of it. The video, posted last week on social media, was created using artificial intelligence and features computer-generated audio of Ossoff claiming to support the shutdown and that he'd 'only seen a farm on Instagram.'

Advanced AI-generated content can now create highly convincing videos of political figures saying things they've never said, or doing things they've never done. The timing of such releases, especially during election cycles, can change public opinion before the deepfake can be discredited. The real danger lies not just in the immediate impact of false content, but in the erosion of trust in politics, creating a 'liar's dividend' where genuine footage can be dismissed as fake.

Specifically, in the political realm, the misuse of deepfakes has accelerated worldwide. According to a 2024 report by the cybersecurity firm Recorded Future, at least 38 countries experienced deepfake incidents targeting public figures within a single year. Most of these cases were linked to elections. Forged audio and video are now regularly attributed to political candidates, eroding public trust and manipulating opinions.

The fundamental concern with deepfakes is that their verisimilitude between reality and the manipulated audio, video, or images can mislead individuals into thinking that the content is real. And yet, election after election in 2024 showed that these warnings were overblown or at least premature.

The malicious use of deepfake technology can lead to violations of human rights and freedoms, or even facilitate criminal activities such as financial fraud. However, creating manipulated images can also pose other threats, including those to democratic states and the principles that govern them.

While deepfakes have dominated global headlines as an existential threat to democracy, the evidence from the 2024–2025 “super election cycle,” during which nearly half of the world's population voted, reveals a more complex reality. The much-feared “deepfake election” did not materialise; however, the convergence of cheapfakes, synthetic media, and algorithmic amplification continues to erode public trust in elections, journalism, and institutions.

The technology — which can do everything from streamlining mundane campaign tasks to creating fake images, video or audio — already has been deployed in some national races around the country and has spread far more widely in elections across the globe. Despite its power as a tool to mislead, efforts to regulate it have been piecemeal or delayed, a gap that could have the potential to significantly impact election outcomes.

Now any individual with basic tech skills and a malicious desire and intent to harm, can sow chaos and potentially influence the outcome of an election. The deepfake could be a video of a president postponing the election due to a cybersecurity attack, a candidate accepting a bribe, or an election official stuffing a ballot box.

The observed intensification of the use of deep fakes to intervene in the will-forming process in open societies or to influence election results is of particular importance in the super-election year. Fear, mistrust, and insecurity thus might be the most important effects of deep fakes.

From the rapid spread of misinformation to the rise of deepfake technology, AI is reshaping democratic processes with significant impact on democracy.

It also examines various deepfake detection methods, legal regulations, and policy measures that can lower the risk of deepfake-related scandals in elections.

Policymakers have expressed concern that deepfakes could mislead voters, but prior research has found that such videos have minimal effects. Our findings suggest that even if deepfakes are not themselves persuasive, information about deepfakes can nevertheless be weaponized to dismiss real political videos.

The EU AI Act, effective 2024, classifies deepfakes as high-risk AI and requires labeling for synthetic content, including political uses, to mitigate election interference risks while balancing innovation.