UK’s independent fact-checking charity, Full Fact. has warned that UK democracy is under “immense and increasing strain” due to the rapid rise of artificial intelligence-driven misinformation, with just 3% of the public able to easily tell whether online videos are real or AI-generated.

According to UK’s independent fact-checking charity there has been a sharp rise in AI-assisted content over the past year, making information easier to manipulate and significantly harder for voters to trust. The organisation is calling for urgent government intervention, including tighter regulation of so-called “political deepfakes” digitally created or altered images, videos and audio designed to mislead the public.

Full Fact cautioned that many people are finding it difficult to discern fact from fiction in an increasingly synthetic information environment, especially during politically sensitive times when false information can spread quickly before corrections are made, as voters in Makerfield get ready for a crucial by-election that could affect the course of UK politics.

According to the report, the problem is no longer restricted to individual misleading claims but rather represents a systemic change in the way information is created, disseminated, and consumed online, with artificial intelligence increasingly speeding up the volume and complexity of distorted content.

Furthermore, a recent study highlights the expanding reach of AI-generated political content on social media platforms by revealing that about 16.5 million UK people, or roughly one in three, were exposed to political deepfakes in the month preceding May’s English local elections.

According to Full Fact, electoral legislation has not kept up with modern campaigning strategies, and voters must be fully informed about who is endorsing political advertising and when AI-generated content is being used in political messaging. It underlined that transparency is now a crucial protection, particularly since it’s becoming harder to distinguish authentic video from deepfakes.

The charity argues that the government “missed the chance to strengthen these safeguards before the recent May elections and the Makerfield by-election,” but still has an opportunity to act before the next general election. It warned that failure to do so could undermine confidence in democratic processes at a time when trust is already fragile.

Chris Morris, chief executive of Full Fact, warned that UK risks falling behind other democracies in tackling the issue.

“Decisions taken now will shape whether future elections are defined by confusion and mistrust, or by clarity, confidence and democratic resilience.”
Chris Morris

He noted that other countries are implementing more integrated frameworks centred on openness, preparation, and accountability, whilst UK’s approach remains too reactive and fragmented. According to Morris, this inconsistency exposes democratic regimes to quickly expanding technology dangers, which are already changing political communication.

Growing Concern Over AI-Generated Misinformation

Furthermore, a YouGov study conducted for Full Fact revealed that 80% of respondents are concerned about political misinformation, while only 3% believe it is simple to tell if web videos are genuine or AI-generated.

Nearly half of the public (48%) said misinformation has already eroded their trust in institutions such as government and Parliament, and two-thirds (66%) believe ministers are doing too little to combat AI-generated false news. The findings point to a growing gap between public expectations and present regulatory actions.

The report also highlights that misleading use of real information is now more common than outright fabrication, with crime, immigration and the economy among the policy areas most frequently linked to misinformation. Analysts say this trend makes fact-checking more difficult, as it often involves selective editing, contextual distortion or subtle manipulation rather than obvious falsehoods.

According to Full Fact, the scale of AI engagement in misinformation has significantly increased. In November 2024, it suspected AI involvement in four of its fact-checking investigations; by October 2025, the number had increased to at least 27. The organization cautioned that this trend signals a larger increase in the use of generative technologies for political messaging and false content generation.

An internal review also discovered that AI-generated summaries and tools were occasionally repeating discredited claims, creating facts about world events, failing to identify synthetic media, and even producing contradicting answers to the same questions.

Full Fact cautioned that such disparities could erode public trust in digital information systems.

Moreover, online information is “becoming harder to trust, more challenging to navigate and all too easy to manipulate,” raising serious concerns for electoral integrity, public debate and institutional confidence.

Full Fact is urging ministers to introduce stronger legal protections against political deepfakes and to update electoral laws to reflect the realities of AI-driven campaigning. It argues that current frameworks were designed for a pre-AI information environment and are no longer sufficient to protect voters.

The charity also recommends giving online platforms a legal duty to support media and political literacy, alongside expanded powers for the Electoral Commission to investigate misinformation and maintain a public archive of political advertising.