Key facts
- One Nation used an AI-generated video with racist tropes at its Victorian campaign launch.
- The video depicted a man appearing African with a machete, a Middle Eastern man with an explosive vest, and a South Asian man discussing "mass migration".
- The video also included a scene of a young man "coming out" as conservative to his leftwing family, similar to an Israeli campaign video.
- The Australian Communications and Media Authority (ACMA) cannot act on the content as it is not broadcast.
- Electoral commissions can only regulate misinformation about the electoral process, not general lies or distasteful content.
- Senator David Pocock has petitioned the government to ban deepfakes in elections.
- The AEC stated AI makes it harder to distinguish authentic from manipulated content during elections.
One Nation's recent use of an AI-generated campaign video featuring racist stereotypes at its Victorian launch has ignited debate about the regulation of artificial intelligence in politics. The video depicted individuals appearing to be of African, Middle Eastern, and South Asian descent in negative and inflammatory ways, alongside a scene mimicking an Israeli campaign video.
The clip, reportedly played at the party's Victorian campaign launch, has drawn criticism for its reliance on tired and offensive tropes. While some argue for ignoring such content to deny the party attention, the broader implications for online politicking and the spread of AI-generated 'slop' are significant.
Australian watchdog bodies currently lack the power to effectively regulate such content. The Australian Communications and Media Authority (ACMA) cannot take action because the video has not been promoted in broadcast forms, and electoral commissions are limited to regulating misinformation about the electoral process itself, not general falsehoods or distasteful material.
Calls for government intervention are growing, with Senator David Pocock advocating for a ban on deepfakes in election campaigns. The parliament's joint committee on AI has also heard testimony that action is overdue, with the Australian Electoral Commission (AEC) warning that AI makes it harder for the public to distinguish authentic content from manipulated material, potentially compromising election integrity.
Professor Terry Flew, from the Centre for AI Trust and Governance, recommended prohibiting AI-generated deepfakes during election periods, emphasizing the risk of further eroding public trust and degrading the information ecosystem. He noted that while the One Nation video may not deceive voters about allegiances, it opens the door to discussions about the appropriateness of such content in campaigns, as AI lowers the barrier for creating inflammatory videos.