Deepfake porn: a rising tide of misogyny

Until recently, the creation of a convincing "deepfake" – a photo or video of an individual that has been digitally manipulated but looks real – required considerable expertise, hundreds of images and massive computing power, said Arwa Mahdawi in The Guardian. 

"Now you just need a couple of photos of someone's face and a phone app." It has become a phenomenon, of a sinister kind: there are now thousands of sites devoted to deepfake pornography. The victims are almost all women, and the latest of them is Taylor Swift. 

Deepfake porn images of the singer were posted online last week, and then spread rapidly across social media – at one stage forcing X/Twitter to block all searches for her name. "If there is a silver lining to this sordid situation", it's that Swift and her millions of followers are a force to be reckoned with. If anyone can get governments and Big Tech to take deepfake porn seriously, it's surely them.

The Week

Escape your echo chamber. Get the facts behind the news, plus analysis from multiple perspectives.

SUBSCRIBE & SAVE

Sign up for The Week's Free Newsletters

From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.

From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.

Sign up

'Bailing out the ocean with a bucket'

As yet, the US has no federal laws against the creation or sharing of deepfake images, said Imran Rahman-Jones on BBC News, though the problem is getting much bigger, very fast. One recent study found an almost six-fold increase in the creation of fake images since 2019, driven by advances in artificial intelligence. In the UK, the new Online Safety Act makes the sharing of deepfake pornography a criminal offence. 

But the rules are hard to enforce, said Sarah Owen in The Independent. Regulators have to distinguish between innocuous fakes (satire, memes or parody) and malign ones (sexual humiliation, misinformation and fraud), and they have to go after the individuals who share the latter, which is like "bailing out the ocean with a bucket".

The law should be changed to crack down on any company involved in unlawful deepfakes; the AI firms whose software makes it possible and the websites that host abusive images should be "liable for the harm they cause".

'Deepfake trickery disrupts elections'

Although it attracted global attention, the Swift episode was arguably not the week's most important deepfake story, said Hugo Rifkind in The Times. That "honour" goes to an automated phone call, made to Democrat voters in New Hampshire, featuring the faked voice of President Biden urging them not to vote in the state primary election. 

Deepfake trickery has the potential to disrupt this year's election campaigns, in both Britain and the US. It's easy to imagine a convincing deepfake audio recording of Keir Starmer in his lawyer days, expressing sympathy for terror suspects. Or one of Rishi Sunak promising to sell off the NHS to private-healthcare companies. Voters must be vigilant: they'll have to learn that they can no longer trust everything they see and hear.