UK political deepfakes are drawing new scrutiny after investigators uncovered overseas pages spreading AI-generated falsehoods about British politics. Researchers say some of the pages were operated from Vietnam while presenting themselves as UK news outlets. The content included fake images, fabricated stories, and manipulated videos involving prominent politicians. The activity has raised fresh concerns ahead of elections in Wales and Scotland in May.
The issue is not limited to one party or one figure. Fake content has featured politicians including Nigel Farage, Boris Johnson, Rishi Sunak, Keir Starmer, and several Welsh leaders. Some posts portrayed politicians in dramatic confrontations or false personal scenarios. Others falsely showed endorsements, arrests, illnesses, or compromising behavior. Investigators found that some removed pages were quickly replaced by new ones.
Experts say the model appears built for virality rather than ideology alone. The pages often post sensational images designed to attract attention and engagement. Some also mixed real news with false claims, making detection harder for casual users. That combination can blur the line between entertainment, misinformation, and political manipulation.
Investigators Trace Pages to Overseas Operators
Researchers found several Facebook pages using names that suggested they were based in Britain. Later, platform transparency tools showed that many were actually run from Vietnam. These pages had thousands of followers and often posted the same or nearly identical material. That pattern suggested coordinated content production rather than isolated user activity.
Some of the fake posts relied on AI-generated images. Others used altered videos to create false political scenes. In one recurring format, multiple politicians were placed into the same invented confrontation. The goal appeared to be emotional reaction, wide sharing, and strong engagement. Researchers described these networks as “content farms” built to go viral.
Meta removed several of the pages after being presented with the findings. The company said it has rules against inauthentic pages and accounts. Still, new pages reportedly appeared throughout the investigation. That suggests enforcement remains difficult when operators can quickly recreate distribution channels.
Researchers also said follower counts may not tell the full story. Some pages showed high follower numbers but low engagement on certain posts. Experts warned that page owners often use bots to amplify visibility and influence algorithms. That can help push fake content into ordinary users’ feeds.
Fake Posts Targeted Politicians and Voters
The false stories covered a wide range of political figures and scenarios. Some posts appeared tailored to attract supporters of Reform UK leader Nigel Farage. In some examples, Farage was falsely shown adopting dogs, donating wealth, or welcoming a child. In others, he was depicted being arrested.
Other politicians were also targeted with invented claims. Keir Starmer was falsely shown being taken ill, sued over election manipulation, or removed as prime minister. Welsh politicians also became subjects of fake videos in recent weeks. These included manipulated clips of Eluned Morgan and Rhun ap Iorwerth in fabricated scenarios.
Several politicians said they had personally been targeted by deepfakes. Alex Davies-Jones said such incidents have become widespread in public life. Her remarks reflected a growing concern among elected officials about reputational damage. Even crude fakes can spread quickly when they fit partisan narratives or emotional instincts.
Not every viewer believed the content. Comment sections often included doubt, irritation, and mockery. Yet some users appeared to accept the stories as real. That mixed reaction highlights the challenge for platforms and regulators. Even low-quality fakes can influence perception if they reach enough people.
Election Officials and Researchers Prepare Responses
The timing has increased anxiety because devolved elections in Wales and Scotland are scheduled for 7 May. The UK has faced earlier warnings about a possible “deepfake election.” A major research institute found no evidence that such content changed the 2024 general election result. However, experts now say barriers to creating AI misinformation have fallen.
That lower barrier matters because fewer technical skills are now required. What once needed specialist software and computing power can now be produced much more easily. Researchers warned that this creates a “trickle-down effect” into devolved and local politics. Smaller campaigns may have fewer resources to respond quickly.
The Electoral Commission is developing software with the Home Office to identify, track, and report deepfakes. Officials say the aim is to help voters recognize misinformation and reduce harm to confidence in campaign conduct. The system may improve detection and reporting. However, experts cautioned it may not stop false content before it spreads.
Some of the recent Welsh examples were described as closer to “shallowfakes” than advanced deepfakes. They were less realistic and apparently made with simpler tools. Even so, researchers said identifying manipulated media is becoming harder. In some cases, teams reportedly spend hours reviewing suspicious clips.
Profit Motives and Platform Pressure
Experts said many of these pages appear driven by profit rather than party affiliation. Viral posts can attract engagement and possibly monetization. That can create incentives to publish whatever draws attention, regardless of truth. Political misinformation becomes commercially useful when outrage increases clicks and shares.
This makes platform enforcement more complicated. False political content may not always come from organized state actors or formal campaigns. It can emerge from low-cost networks seeking advertising revenue or follower growth. That widens the threat landscape for elections and public trust.
The broader concern is not only whether a single fake changes a single vote. Repeated exposure to manipulated content can erode trust in politicians, media, and election systems. That loss of trust can become its own political effect. As UK political deepfakes become easier to create, the pressure on platforms and regulators is likely to intensify.
