A nurse breaking down in tears, claiming she was sacked for posting her views on immigration online. A farmer accusing the Labour government of deliberately destroying British agriculture in favour of imported foreign meat. A man watching migrants arrive on England’s shores, claiming they are handed luxury accommodation on arrival.
These politically charged, divisive videos about Britain have racked up millions of views, but they’re all created using artificial intelligence (AI) and produced by people more than 5,000 miles away.
The page, Life in Britain, has more than 100,000 Facebook followers. Despite all of its videos being AI-generated, none come with warnings from Facebook, owned by parent company Meta, despite The Independent flagging this with the tech firm more than a month ago.
Facebook’s transparency information on the page shows that the page’s managers are based in Sri Lanka, despite producing content aimed at UK audiences.
Experts at Resemble AI, a deepfake and synthetic media detection company, told The Independent the account could be part of an AI-generated influence campaign designed to amplify British political narratives online.
Resemble AI analysed the videos for The Independent and found several signs they had been generated using AI, including unnatural audio and visual flaws such as distorted faces. It assessed the videos as 97 per cent likely to have been AI-generated.
“It’s not uncommon to get social engineering attacks from foreign entities. You get scam calls that don’t originate at all inside the UK. This is very similar,” Zohaib Ahmed, Founder at Resemble AI said.
Facebook told The Independent on 29 May that it was investigating the account after it was brought to their attention, but did not provide a timeline for the investigation or say whether any action would be taken. When asked whether this account was monetised, Facebook did not respond.
Dr Lukasz Olejnik, visiting senior researcher fellow at the Department of War Studies, King’s College London, told The Independent: “The use of AI-generated influence and propaganda content is on the rise. AI-generated video is particularly significant because it can create false visual evidence that is easy to share, emotionally persuasive, and difficult to verify quickly. These tools can be used by individuals, coordinated non-state and state actors”.
Dr Olejnik added: “For the UK, the risk is not only that voters may believe a single false video, but that repeated exposure can weaken trust in institutions, journalism, elections, and foreign policy decisions.
“And in the leadership, including in times of crisis. This creates opportunities for external actors to exploit sensitive political debates, like security, migration, protest, defence policy, and relations with allies.”
The potential ramifications of this kind of online content has not gone unnoticed. Last year, research submitted to the House of Commons Home Affairs Committee showed AI-generated images portraying Muslims as a threat fuelled the Southport riots in the summer of 2024.
Far-right accounts were reportedly spreading manipulated images online that reinforced racist stereotypes, including depictions of Muslims carrying guns or wearing suicide vests.
The London School of Economics study found that social media posts containing these visual depictions of racist conspiracy theories were amplified 30 per cent more than others.
Researchers from LSE warned that the rise of AI-generated material, combined with the way social media algorithms reward divisive content, is accelerating radicalisation and making extremist ideas more visible.
The rise of online disinformation prompted London mayor Sir Sadiq Khan last month to pledge £7m to tackle the issue after saying the capital had been subjected to a “relentless and unprecedented attack of lies and hatred”.
The announcement followed a report by the Greater London Authority (GLA), which found a 200 per cent increase over the past two years in online posts portraying London as a city in decline and increasingly dangerous.
“If I wanted to influence decisions, the easiest way to do so at this point becomes creating generative AI content, creating fake AI accounts on social media platforms. And then you have a distribution channel” Mr Ahmed said.
He added: “Those algorithms will float that content to millions of people for hundreds of thousands of millions of views almost immediately. So it becomes almost a mathematical game.”
Earlier this year, Meta’s independent Oversight Board warned the company should label AI-generated content on its platform more proactively and “more frequently”.
They recommended that Meta invest in stronger detection tools for AI-generated content.
The Department for Science, Innovation and Technology said: “Platforms have legal duties to tackle content that is illegal, including material which amounts to incitement to violence or hatred, and we expect them to take robust action to protect users.”
Facebook has clear guidelines that prohibit hate speech, harmful misinformation, deceptive behaviour and coordinated inauthentic activity, and they claim to take action on content that violate policies.
Earlier this year the tech platform removed several Vietnam-based Facebook pages described as “content farms” after BBC Wales found they were spreading false narratives about British politics using deepfakes.
The Life in Britain Facebook page owners have been contacted for comment.
