By Stuart Kerr, Technology Correspondent, LiveAIWire
AI in elections reached a genuine inflection point during the 2024 UK general election, when researchers at the Oxford Internet Institute identified over 400 pieces of AI-generated political content circulating on UK social media platforms, including synthetic audio clips purporting to be senior politicians, AI-generated images depicting party leaders in fabricated contexts, and automated content farms producing high-volume partisan messaging at a scale no human campaign team could match. None of these were produced by the major parties. Most could not be traced to any identifiable source. And the regulatory framework designed to ensure electoral fairness had no meaningful response to any of them.
The 2024 UK general election was not the first election affected by AI-generated disinformation, LiveAIWire has examined the mechanics of that broader phenomenon in our coverage of AI political deepfakes. What made the UK campaign distinct was not the technology itself but the exposure of just how unprepared British electoral law is for it.
The AI Content Landscape During the Campaign
The AI content documented by researchers during the 2024 UK election campaign fell into several categories with different risk profiles. Synthetic audio, designed to sound like genuine recordings of politicians, carried the highest disinformation risk because audio is inherently difficult for most people to evaluate for authenticity without specialist tools. Fabricated images had lower credibility with sophisticated audiences but circulated widely in WhatsApp groups and community forums where fact-checking resources were not routinely available.
The platforms on which this content spread varied in their responses. Meta’s Facebook and Instagram flagged some synthetic content through their AI detection tools but missed a significant proportion. TikTok, which has a younger and more algorithmically engaged audience, was the platform where AI-generated political content spread fastest with least friction. WhatsApp, as an end-to-end encrypted private messaging platform, was essentially outside the scope of any platform moderation approach. The Electoral Commission acknowledged that its existing imprint requirements for campaign materials did not apply to most AI-generated political content and that its enforcement powers were inadequate for the digital environment.
Why UK Electoral Law Cannot Handle AI in Elections
UK electoral law is built around a framework developed for print and broadcast media. The requirements for campaign material imprints, spending limits, and broadcaster balance obligations all presuppose a media environment in which content is produced and distributed by identifiable parties at costs that campaign finance rules can track. None of these assumptions holds for AI-generated political content, which can be produced at near-zero cost, by actors with no connection to registered campaigns, without any imprint, and distributed through networks that are algorithmically amplified without any spending that counts against electoral limits.
The Law Commission’s review of electoral law, published in 2020, identified the digital media gap as a priority reform area. The Elections Act 2022 made some updates but did not comprehensively address AI-generated campaign content. A cross-party parliamentary working group on AI and elections recommended mandatory AI disclosure labels for political advertising, criminal penalties for distribution of knowingly false synthetic political media, and mandatory platform transparency reporting during election periods. These recommendations have not been implemented, and the next election will face the same regulatory environment as the last one, with more capable and more accessible AI tools available to those who wish to misuse them.
How Other Democracies Compare
The UK’s regulatory position on AI in elections compares unfavourably with several peer democracies. The European Union’s revised Code of Practice on Disinformation requires platforms to label AI-generated political content and report on enforcement during election periods, with significantly more enforceability than equivalent UK commitments. Germany passed specific legislation requiring disclosure of AI-generated political advertising ahead of its 2025 federal election. Taiwan, which faces sophisticated AI disinformation campaigns from hostile state actors, has developed a whole-of-government approach to election integrity that includes specific AI provisions. The International Foundation for Electoral Systems has documented a growing gap between the sophistication of AI disinformation threats and the regulatory responses of most established democracies.
What This Means for You
Voters in any democracy facing AI in elections now need to apply a higher level of scepticism to political content than was necessary a decade ago. Audio recordings of politicians making controversial statements, images depicting party leaders in compromising situations, and high-volume content that appears to come from grassroots movements should all be verified against multiple reliable sources before being shared or allowed to shape voting decisions.
This individual responsibility in the face of AI in elections is real but insufficient. The structural solutions, regulation of AI political content, platform transparency requirements, and adequate enforcement resources for the Electoral Commission, require political decisions that elected representatives have been slow to make. Cross-party consensus on the basic proposition that AI-generated synthetic content used to mislead voters is an electoral integrity threat that requires a legislative response would provide the foundation for action that has so far been absent. For related coverage of how AI is reshaping the wider information environment voters rely on, see LiveAIWire’s analysis of AI in journalism.
The Deeper Democratic Stakes
The deeper question raised by AI in elections is about the relationship between information quality and democratic legitimacy, a tension LiveAIWire has also traced in our coverage of AI conspiracy culture more broadly. Elections are supposed to aggregate genuine preferences based on genuine information. When the information environment is systematically polluted by synthetic content designed to mislead, the aggregation process produces outcomes that do not accurately reflect informed voter preferences. Political scientists have documented relationships between information quality and electoral outcomes that suggest AI in elections is already measurably affecting how people vote.
Addressing this requires not only technical and regulatory responses but investment in the media literacy education that helps voters navigate a synthetic information environment. The Full Fact organisation in the UK has published evidence-based recommendations for election integrity in the AI era that deserve more political attention than they have received. The 2024 election demonstrated the problem clearly enough. The next one should not have to demonstrate it again.
About the Author
Stuart Kerr is Technology Correspondent at LiveAIWire, covering artificial intelligence, cybersecurity, and the social impact of emerging technology. He publishes daily at LiveAIWire.com.