A recent study has revealed a network of numerous Facebook pages based outside of Australia that are generating artificial intelligence (AI) deepfakes featuring local politicians. This foreign operation reportedly includes around 10% of the pages that have been monetized, enabling their operators to gain financially from what researchers describe as a commercially driven “foreign influence campaign.”
The findings illustrate a troubling trend where “AI slopaganda” networks are targeting political systems across the globe on an extensive scale. Reset Tech, the research organization behind the study, has identified over 200 Facebook pages that fabricate content featuring prominent Australian politicians, including Prime Minister Anthony Albanese and opposition leader Angus Taylor, among others, over the last eight months.
A report from Reset Tech has linked this coordinated campaign to a business operating from Sri Lanka. The research indicates that approximately one in ten of the identified pages has been monetized through Facebook’s content creator program, allowing page owners to profit directly from their posts.
“This is not foreign interference in the traditional sense,” stated Rys Farthing, the chief research officer at Reset Tech. “What we are witnessing is foreign influence driven by commercial interests.”
Many of the deepfake videos target politicians in brief clips shared via Facebook Reels, where they are misleadingly depicted making inflammatory statements. The content frequently promotes anti-migrant, anti-Muslim, or nationalistic narratives, while also encouraging viewer engagement through polls. Additionally, there is a recurring trend of using sexualized imagery of female politicians.
“While some content relates to current events, much of it seems random,” Dr. Farthing noted. “They test various posts to see which ones attract attention and revenue, then replicate that style.”
Senator Fatima Payman, who transitioned from the Labor party to sit as an independent, was a victim of a page known as Australia Info Live. In an AI-generated video, she was falsely depicted as saying, “Please don’t deport me back to Afghanistan. It’s not safe for me there. Vote no to support me.” The post garnered numerous likes and comments without any indication that it was fabricated.
Another independent senator, David Pocock, was targeted by Australia Life & News, where he was misleadingly portrayed as having “joined forces” with One Nation. Pocock expressed his concern about deepfakes and their potential to undermine democratic processes, emphasizing that people often trust what they see.
Pages like I Love Australia, Life in Australia, and Beautiful Australia are typically designed to attract a wide audience. Reset Tech identified a common strategy of drawing in viewers with AI-generated images of animals or attractive figures before shifting to political content, although many of these pages are politically driven from the start.
“This is not an issue confined to domestic actors,” Dr. Farthing remarked. “It represents a foreign manipulation of Australian political dialogue.”
Home Affairs Minister Tony Burke, who has also been a victim of deepfakes from this network, stressed the importance of verifying information sources. He noted that during the last election, he encountered photographs of himself alongside individuals he had never met, and more recently, he has seen videos of his colleagues making statements they never actually made.
Reset Tech’s investigation indicated that the same Sri Lankan network is also targeting various countries including France, Italy, the UK, the US, Canada, and Mexico, employing similar tactics. It is estimated that this network operates 1,864 coordinated Facebook pages globally, sharing a common branding and amassing a total of 5.9 million followers.
“We aimed to highlight that this issue isn’t confined to a single perpetrator or a few bad actors misusing the platform,” Dr. Farthing stated. Instead, the research illustrates that “AI slopaganda” networks are systematically targeting political landscapes on an extensive level.
Reset Tech’s findings indicated that instances of political deepfakes on Facebook are rarely addressed in line with the platform’s policies. The report noted that only one-third of the posts received AI labeling, and just one in thirty-three posts were removed. “We were able to uncover and identify these networks with relative ease,” Dr. Farthing pointed out. “If we can do it, I struggle to understand how these billion-dollar companies cannot.”
According to Reset Tech, the network targeting Australia is violating multiple Facebook policies meant to combat inauthentic behavior, hate speech, and AI deepfakes. Dr. Farthing concluded that social media platforms are inadvertently promoting the creation of AI slop by failing to contain it. “The decisions they’ve made have facilitated their monetization system,” she remarked.
Senator Pocock added that it is becoming increasingly clear that addressing this issue is not aligned with the interests of social media companies. “Their business model thrives on outrage, maximizing screen time, and generating clicks,” he noted. The Australian Broadcasting Corporation has reached out to Meta, Facebook’s parent company, for their response to the Reset Tech report.
