Warning AI could ‘steal next election’

Greens senator David Shoebridge has warned that sophisticated deepfake videos made using generative AI has the power to “steal the next federal election”.

The warning comes after the final report into a Senate probe into adopting artificial intelligence was delayed on Thursday, with the report now due to be published by November 26.

However, an interim report to be released in the coming weeks will focus on the potential of AI-generated deepfake political material and it’s effect on elections, with the US election slated for November 5, and an Australian federal election slated for next year.

Senator Shoebridge said Australia’s laws targeting AI-generated deepfake content were out of date, and said more powers needed to be given to the Electoral Commission to prevent deceptive content.

“Everyone who cares about democracy should be concerned about the potential for AI deepfakes being used to steal the next Federal election,” he told NewsWire.

“We are seeing this play out in real time in the US presidential campaign with deepfakes and altered videos widely circulating.

“That should really be spurring other MPs and Senators to join us and start protecting our democracy.”

During a May hearing, Australian Electoral Commissioner Tom Rogers said the watchdog did not have the “legislative tools” or the “technical capacity” to detect AI-generated content, and conceded Australian voters were not immune to the fake content.

Earlier this month, independent senator David Pocock shared a generative AI video of Prime Minister Anthony Albanese announcing a ban on gambling ads, highlighting the ease of how the videos are produced.

“I’m concerned we’re not seeing the urgency required to protect our democracy from generative AI,” he said.

However, committee chair and senator Tony Sheldon said any potential laws needed to “strike the right balance” between upholding free speech and political expression, while protecting vulnerable people suspectable to being influenced by the doctored content.

“Political deepfake laws, like any reforms relating to political expression, need to be carefully considered, widely consulted on, and receive broad support across the political spectrum,” he said.

“If we aren’t careful with how we regulate political deepfakes we risk achieving the very outcome we’re trying to avoid – reducing public faith in our democratic institutions.”

Currently Labor has introduced a Bill to combat misinformation and disinformation,

and will cover fake political content, and allows the Australian Communications and Media Authority to fine companies up to 5 per cent of their global turnover.

It also allows compels the regulator to compel companies for information on how they are preventing mis and dis information, however ACMA cannot force platforms to remove content.

Although the Greens have yet to announce their support for the Bill, the Coalition have; lashed it over concerns it would impede free speech.