Australian politicians deepfaked by foreign influence network on Facebook, researchers say

Dozens of Australian politicians have been deepfaked by a "foreign influence" network operating on Facebook, and researchers say there are many more just like it.

Research organisation Reset Tech has tracked more than 200 Facebook pages impersonating Prime Minister Anthony Albanese, opposition leader Angus Taylor, One Nation's Pauline Hanson, and many more political figures over the past eight months.

A report from the group has revealed the seemingly co-ordinated campaign and linked it back to an influencer business in Sri Lanka.

It found roughly 10 per cent of the network's pages have been monetised by Facebook, as part of the platform's in-house scheme to pay content creators, allowing a page's owner to profit directly.

"This isn't foreign interference as we traditionally think of it," Rys Farthing, chief research officer for Reset Tech, said.

Researchers found most targets were impersonated in short clips on Facebook Reels, in which they were falsely shown to be saying something inflammatory.

The content often centred on anti-migrant, anti-Muslim, or nationalistic themes, and baited users into participating in polls.

Another common pattern was the use of sexualised images and videos of female politicians.

"Sometimes it is directly related to news events, but basically … it's just throwing spaghetti at the wall," Dr Farthing said.

"They work out which [posts] get them the clicks, which ones bring the money in, and then they create more like that again."

Senator Fatima Payman, who left Labor to sit on the crossbench, was among those targeted by a page called Australia Info Live.

Senator Payman, a Muslim, was falsely shown in an AI-generated clip to say: "Please don't deport me back to Afghanistan. It's not safe for me there. Vote no to support me."

The post has hundreds of likes and comments, and no apparent label to indicate it was false or AI-generated.

Fellow independent senator David Pocock was targeted by a page called Australia Life & News, in a clip where he was falsely shown to say he had "joined forces" with One Nation.

Senator Pocock told the ABC he was not aware of the specific clip, although his team had been sent "a bunch of deepfakes of me over the last few months".

"I'm concerned deepfakes really undermine our democratic system," he said.

"As humans, when we see someone say something, we generally believe that it's them saying it".

With names such as I Love Australia, Life in Australia, or Beautiful Australia, the pages are often designed for mass appeal.

Reset Tech said one of the common tactics was to lure audiences in with AI-generated images of animals, attractive women, or posts praising Australia's beauty.

Later, the page takes a political turn, although many pages are political from the outset.

"There's nothing domestic about this," Dr Farthing said.

Home Affairs Minister Tony Burke, who has been deepfaked by the same network, said it was more important than ever that people checked their sources.

"Last election for the first time I saw photos of myself online with people I had never met," he said.

"In more recent times, I have seen videos … and interviews of some of my colleagues saying things they have never said."

Reset Tech found the same Sri Lanka-based network was also targeting France, Italy, the UK, the US, Canada and Mexico, among other locations, using the same playbook.

In total, it was believed to operate 1,864 coordinated Facebook pages globally, with mostly identical branding, and a combined follower base of 5.9 million users.

"We really wanted to highlight that this wasn't just a unique issue with one bad actor or a couple of bad actors misusing a platform system," Dr Farthing said.

Rather, she said, the research showed "AI slopaganda" networks were targeting political systems on a "massive industrial scale".

'Outrage, eyeballs and clicks': The monetisation system fuelling AI slop

Reset Tech found relatively few instances where political AI slop on Facebook had been appropriately dealt with, in accordance with the platform's own policies.

Only one in three of the posts received AI labelling, and one in 33 of the posts had been deleted, the report stated.

"We have a handful of researchers, and we were able to find this and detect all of these networks with ease," Dr Farthing said.

"If we can do it, I don't quite understand how these companies that have billion-dollar revenues can't do it."

Reset Tech said the network targeting Australia was in breach of several Facebook policies designed to target "inauthentic behaviour networks", as well as hate speech and AI deepfakes.

Dr Farthing said ultimately, platforms were incentivising the creation of AI slop and failing to contain it.

"[The platforms] have made all of the decisions that enable their monetisation system to work, right?"

"And this is the result."

Senator Pocock said it was increasingly clear tackling the issue was not in the interests of social media companies.

The ABC has contacted Meta, Facebook's parent company, for a response to Reset's Tech report.