One Nation’s new AI-generated campaign video is a new low – even for them.
The video – reportedly played at the party’s Victorian campaign launch over the weekend – is an AI-generated collection of racist tropes, depicting a man who appears to be African brandishing a machete, an apparently Middle Eastern man with an explosive vest, a South Asian man in a turban talking about “mass migration” and an Indigenous elder talking about people receiving “money … on the basis of need, not identity”.
But the inflammatory clip says more about the wild west of online politicking created by cheap, unrestricted access to AI videos, and why governments need to empower watchdog bodies to do more about it.
The video was barely mentioned in most media reporting, which otherwise largely focused on the rousing reception Pauline Hanson received from her supporters, the music that was played over loudspeakers and the Trump-esque pledge from candidate Mark Nicholson – the producer of the “Please Explain” cartoon series – to jail Daniel Andrews and Jacinta Allan. Guardian Australia, previously barred from One Nation events, did not attempt to attend Saturday’s launch.
Coverage has focused on the first part of the video, an AI-generated scene of a young man “coming out” as a conservative to his leftwing family. As progressive campaign group GetUp observed, it’s a scene cribbing from a similar Israeli campaign video from Benjamin Netanyahu.
It could also be pointed out how tired the gags are. Blue-haired hipsters with nose rings, eating mung beans? Revolutionary. Were the jokes generated by ChatGPT as well?
There’s an argument for ignoring it entirely, so as to not give it the attention and free advertising the party obviously wants. It is rage bait, seeking to draw a response and then complain about the response as further proof that One Nation is being “picked on” or is “riling up the lefties”. The response of a Hobart cafe frequented by Hanson’s daughter Lee pledging to donate takings from One Nation customers to the Greens – and accompanied by predictable “genuine hurt” from the younger Hanson – was entirely on script.
But for all the bad taste, there’s little anyone can do about it right now.
The Australian Communications and Media Authority cannot take action, because its powers don’t extend to this kind of content, and the video thus far seems to only exist online – not promoted in broadcast forms. Electoral commissions only have powers to regulate misinformation about the electoral process – not lies or distasteful content.
The federal Labor government has so far shrugged off growing calls from inside and outside parliament to ban the use of AI-generated content in election periods. As the government works to legislate around a raft of AI-related harms this year, the time would be ripe to deal with this long-ignored issue.
News Corp reported a One Nation spokesperson saying they were preparing “another version for broadcast”. That could technically put it inside Acma’s remit, but the tone of the ad is unlikely to breach current broadcasting rules, one industry source said.
So, then, is there something to the idea of clamping down on AI-generated content more broadly? Senator David Pocock thinks so: he’s been pushing for two years for the government to address generative AI around election campaigns, with 35,000 people signing his online petition calling for Labor to “legislate a ban on the use of deepfakes in elections before the next federal election”.
This week, the parliament’s joint committee on AI heard from witnesses saying action was overdue. The AEC said in a submission to the inquiry that AI makes it “more difficult for people to distinguish between authentic and manipulated content, increase[s] confusion about authoritative sources of information, and create[s] challenges for [electoral agencies]”.
“Inauthentic content or deepfakes circulating during the voting period, for instance, may well compromise the actual or perceived integrity of an election.”
These are not new concerns. A parliamentary report from October 2024 warned of AI “being used to adversely influence or undermine electoral and democratic processes” and “to deceive the public, create or exacerbate bias or influence public opinion or sentiment corrosively”.
Of course, the One Nation video is a different beast. No voter is being tricked into thinking a leftwing family has really changed its allegiances. But it should open the door to a wider discussion about whether this kind of content is appropriate for an election campaign.
AI lowers the barrier for entry for making such inflammatory videos. Normally, it would be harder – and more expensive – to hire a film crew and actors willing to act out such rhetoric.
In 2024, the Macquarie Dictionary named “enshittification” as its word of the year, defining the term as: “The gradual deterioration of a service or product brought about by a reduction in the quality of service provided, especially of an online platform, and as a consequence of profit-seeking.” Even casual users of Facebook, X, email platforms, Google search and basically any online marketplace would be familiar with the term.
Australian politics, like many parts of the world economy, is enduring what I’m calling “ensloppification”: the permeation of AI-generated slop into every crevice of society – and politics is not immune.
This week we have seen conservative campaign group Advance’s own Victorian AI ad, picturing a Labor election win as Daniel Andrews pops champagne corks and leads a conga line. Liberal MP Jason Wood was factchecked by AAP for posting “a fake image of Anthony Albanese arriving at a deserted airport”, while AAP has also debunked numerous AI-generated political interview-style videos purportedly shot outside Parliament House.
Prof Terry Flew, of the Centre for AI Trust and Governance, told the inquiry on Wednesday that Australia had no single agency to address all misinformation risks, but recommended that government consider prohibiting AI-generated deepfakes during election periods.
“There are real risks that AI-generated misinformation, as well as what has come to be known as ‘AI slop’, will further erode public trust in institutions and degrade the wider information ecosystem,” Flew said.
“[AI-generated deepfakes go] beyond the legitimate bounds of public campaigning and prove to be genuinely threatening to the individuals and groups who are the targets of them.”
-
Josh Butler is a Guardian Australia political reporter and chief of staff in Canberra
Disclaimer : This story is auto aggregated by a computer programme and has not been created or edited by DOWNTHENEWS. Publisher: theguardian.com
