91av

LAST week, UK streaming service ITVX unleashed a world first. It came in the form of a sketch show in which some of the world’s biggest stars, from actor Idris Elba to climate campaigner Greta Thunberg, are reimagined as squabbling neighbours, drawing battle lines over communal gardens and Christmas decorations.

More significant than these quarrels between famous faces, though, is how those faces were created. Deep Fake Neighbour Wars takes the voices and bodies of impressionists, then marries them with digital versions of celebrity visages. Welcome to the first ever deepfake comedy.

I have started watching and I won’t pretend it is must-see TV: the impersonations are patchy and the deepfakes, though technically solid, are somewhat uncanny. The level of realism also means the show has to pull its satirical punches, lest any litigious noses be put out of joint.

But the success of the underlying technology is clear, opening a can of worms about the future of deepfakes in television and film. It is a future that I believe can be positive, if we finally wrestle with the societal maladies that have led to deepfakes’ misuse.

Entertainment companies are already using deepfakes for positive purposes. Last year, the makers of the film Fall used the tech , avoiding costly reshoots and securing a US rating for younger audiences. Deepfakes have tackled language barriers in , such as in ads for a malaria charity in which speaks in nine languages. They have helped conceal the identities of people in documentaries like Hong Kong’s Fight for Freedom.

But there are real concerns about how they could be deployed, even when setting aside malicious applications, such as in deepfake porn that misappropriates facial images of well-known people. It is also easy to imagine, for instance, the tech taking work away from actors. Questions remain over how deepfakes affect performers’ image rights – if they have them at all. “Image rights don’t really exist in UK law. And even where they do, like in some US states, the devil is in the detail,” says at Newcastle University, UK.

What is to stop a firm using a person’s face without consent, provided it hasn’t breached copyright? While some public figures have licensed their image rights for “legitimate” deepfakes, such as and , ethical and legal norms around are far from established.

Things get even murkier when it comes to the dead. In 2021, the late food writer in the documentary Roadrunner, prompting from his friends and family, particularly as the film doesn’t disclose that it uses this technology (unlike Deep Fake Neighbour Wars).

I can understand those who feel deepfakes’ benefits will always be overshadowed by abuse. But we can’t uninvent this technology and our attempts to control its use have lagged behind its growing sophistication. While we must lobby for laws around AI to keep pace with these developments, it is even more important we talk about how to use them ethically.

The most worrying applications all stem from deeper issues: the erosion of employment and privacy rights; hatred of women; a lack of media literacy; and of accountability for putting out misinformation. To tackle harmful deepfakes, we must remember that at the heart of this problem is a neutral tool that simply reflects humanity’s evils back at us.

“It shines a light on inadequacies that are nothing to do with AI,” says at the University of Surrey, UK. “So how do we solve that?” If you tune in to Deep Fake Neighbour Wars, this is the question you should ask yourself.

Bethan Ackerley is New Scientist’s television columnist and a subeditor