‘You don’t need to be a celebrity for this to happen’: Minister urges Australians to seek help if targets of deepfakes

Australia is looking to step up penalties for online platforms that ignore takedown orders for intimate pictures, including AI “deepfakes” like the viral images of Taylor Swift.

Australia is looking to step up penalties for online platforms who ignore take down orders for non-consensual intimate images, including AI “deepfakes,” as the federal government warns tech giants their “cavalier” approach to safety “stops now”.

The scourge of AI generated explicit images has exploded into the mainstream since the internet was flooded with false pornographic images of singing superstar Taylor Swift, but Communications Minister Michelle Rowland says celebrities are not the only victims of the “deeply distressing” crime.

Unlike Swift in the US, Australians can report any abuse to the eSafety Commissioner, which has the power to compel a social media platform to remove a post within 24 hours.

To date the Commissioner has a 90 per cent success rate, but Ms Rowland said the government’s “main challenge” was raising awareness about the help available.

“There’s the ability for people to report non-consensual sharing of those intimate images by going to esafety.gov.au,” she said.

Ms Rowland confirmed that when in Australia, even Swift herself could report abuse to the Commissioner.

Communications Minister Michelle Rowland says Australians can get help at esafety.gov.au if targeted online. Picture: Supplied

“The Taylor Swift incident really heightened awareness, but you don’t need to be a celebrity for this to happen,” she said.

“All Australians should know … if they have intimate images shared of them, they have opportunities for redress.”

Failure to comply with a notice to remove content can result in fines of about $700,000.

Ms Rowland said a review of Australia’s Online Safety Act brought forward to this year would have broad scope to consider gaps in the current legislation, including whether current penalties are “adequate”.

“(The review will be) looking at the amount of fines and whether they are a sufficient deterrent, or an incentive for platforms to act quickly,” she said.

Work is also underway to update Australia’s basic online safety expectations for digital platforms, which Ms Rowland said would particularly aim to capture the new challenges of generative AI.

“The industry cannot be so cavalier in designing algorithms, apps and other products that do not promote safety,” she said.

“As far as this government is concerned, it’s unacceptable and it stops now.”

Recent viral AI deepfakes of Taylor Swift have renewed the focus on online safety. Picture: Michael Tran / AFP

After being flooded with complaints from Swift’s fans last month social media platform X blocked users from being able to search the singer’s name in order to limit the spread of highly explicit deepfakes already viewed millions of times.

Ms Rowland said platforms knew their consumers “better than anyone,” including governments and regulators, and therefor had a “responsibility” to use that information to keep people safe online.

“Requiring platforms to take certain principles into account when they are designing their products … is one of the elements in the basic online safety expectations,” she said.

Australia’s safety legislation had already accounted for the advent of “deepfake” images, putting the nation ahead of counterparts like the US.

However Ms Rowland said with the pace of emerging technologies now the task of safety was a “job never done” that must be constantly reviewed.

Do you have a story for The Daily Telegraph? Email [e-mail]