The British government has announced plans to criminalize the creation of sexually explicit deepfakes, with perpetrators facing up to two years behind bars if found guilty.
It is already an offense to share or threaten to share intimate images, including deepfakes, following amendments to the Sexual Offences Act 2003 that were made by the Online Safety Act 2023.
However, the government plans to additionally criminalize the creation of such deepfakes, as well as the taking or recording of an intimate photograph or film without consent – whether the intent is sexual gratification or “to cause alarm, distress, or humiliation.”
“It is unacceptable that one in three women have been victims of online abuse. This demeaning and disgusting form of chauvinism must not become normalized, and as part of our Plan for Change we are bearing down on violence against women – whatever form it takes,” said victims minister, Alex Davies-Jones.
“These new offenses will help prevent people being victimized online. We are putting offenders on notice – they will face the full force of the law.”
Read more on deepfakes: One Deepfake Digital Identity Attack Strikes Every Five Minutes
Sexually explicit, AI-generated images are a growing problem – often used by cyber-bullies and online extortionists.
As far back as 2023, the FBI warned of a surge in complaints about digital sextortion attempts, with malicious actors manipulating benign images or videos of victims which they scraped from social media and other publicly facing sites.
The recent emergence of generative AI tools has further lowered the barrier to entry for would-be sextortionists, with the technology both free to use and the output increasingly realistic.
However, experts were not convinced the proposed law would discourage malfeasance.
ESET global cybersecurity advisor, Jake Moore, argued that police may struggle to root out offenders.
“Deepfake technology is improving at a tremendous rate and the police have a difficult time ahead of them as deepfakes can be incredibly difficult to assign a creator,” he added. “This is because many unethical AI services remain under the radar either on the dark web or as apps within encrypted communication channels.”