Deepfake nudes and “revenge porn” must be removed from the internet within 48 hours or technology firms risk being blocked in the UK, Keir Starmer has said, calling it a “national emergency” that the government must confront.
Companies could be fined millions or even blocked altogether if they allow the images to spread or be reposted after victims give notice.
Amendments will be made to the crime and policing bill to also regulate AI chatbots such as X’s Grok, which generated nonconsensual images of women in bikinis or in compromising positions until the government threatened action against Elon Musk’s company.
Writing for the Guardian, Starmer said: “The burden of tackling abuse must no longer fall on victims. It must fall on perpetrators and on the companies that enable harm.”
The prime minister said that institutional misogyny being “woven into the fabric of our institutions” meant the problem had not been taken seriously enough. “Too often, misogyny is excused, minimised or ignored. The arguments of women are dismissed as exaggerated or ‘one-offs’. That culture creates permission,” Starmer wrote.
Government sources said they expected to give the new powers to Ofcom to enforce by the summer and companies will be legally required to remove this content no more than 48 hours after it is flagged to them.
Platforms including social media companies and pornography sites that fail to act could face fines of up to 10% of their qualifying worldwide revenue or having their services blocked in the UK.
Victims would be able to flag the images either directly with tech firms or with Ofcom – which would trigger an alert across multiple platforms, according to the Department for Science, Innovation and Technology.
Ofcom would be responsible for enforcing the ban on the images, with the aim to remove the onus on victims to need to report the same image potentially thousands of times as it is continually reposted.
The media regulator will be told to explore ways for “revenge porn” images to have a digital watermark to allow them to be automatically flagged every time they are reposted.
Internet providers will also be given new guidance on how to block hosting for rogue sites who specialise in hosting nonconsensual real or AI-generated explicit content.
The Grok “nudification” tool that sparked an outcry in early January saw ministers threaten to ban X if it did not act. About 6,000 bikini demands were being made to the chatbot every hour, according to analysis conducted for the Guardian, with many requests made to create images of women bent over or wearing just dental floss.
But there has also been a rise in recent years of nonconsensual real or deepfake images being used to blackmail young women and men, which charities have linked to a number of suicides.
Starmer said the horror stories of women and girls who saw intimate images spread across the internet was “the type of story that, as a parent, makes your heart drop to your stomach”.
“Too often, those victims have been left to fight alone – chasing action site to site, reporting the same material again and again, only to see it reappear elsewhere hours later,” the prime minister said. “That is not justice. It is failure. And it is sending a message to the young people of this country that women and girls are a commodity to be used and shared.”
Creating or sharing nonconsensual intimate images will also become a “priority offence” under the Online Safety Act, giving it the same level of seriousness as child abuse images or terrorism. The law does not require platforms to independently identify nonconsensual intimate imagery, but just to remove these images when they are flagged.
Google, Meta, X and others already do this for child sexual abuse content through a process called hash matching – which assigns videos a unique digital signature that can be matched against databases of abusive content. And while the 48-hour timeline is brief, India has recently mandated that social media companies remove some deepfake content in three hours.
“I think that 48 hours is certainly possible, to be honest with you,” said Anne Craanen, who researches online misogyny at the Institute for Strategic Dialogue.
“The problem is, it may not necessarily incentivise a quicker response rate from companies. But 48 hours is longer than the timeframe for the removal of other types of content, such as terrorist content in the EU.”
Craanen added that there are already existing initiatives to use hash matching to protect victims of intimate abuse; although it can be challenging getting different tech platforms to coordinate among themselves, so that an abusive video uploaded to Facebook, for example, is automatically detected on Reddit.
Hash matching is not a perfect technology, Craanen emphasised, and can be circumvented. Terrorist groups, for example, often add emojis or small alterations to videos already hashed as terrorist content, thus making them unrecognisable to hash matching systems.
The advent of AI tools and AI deepfakes will make this problem worse, allowing nonconsensual intimate images and other content to be quickly altered and spread around the internet, Craanen said, evading attempts to quickly detect it with tools such as hash matching. In a moment like the January Grok bikinification crisis, some abuse might be impossible to reign in.
While the law appears to apply to all tech platforms, including “rogue websites” that fall outside the reach of the Online Safety Act, there are questions as to how it could apply on encrypted messaging services such as WhatsApp and Signal.
In his article, Starmer said he was also determined to challenge misogyny in government and in politics, after several weeks where the prime minister has faced criticism for the appointment of Peter Mandelson as US ambassador despite knowing about his friendship with the disgraced financier and child sex offender Jeffery Epstein. Mandelson was fired after new revelations about the closeness of their friendship.
The prime minister is also facing a row over the appointment of a new cabinet secretary, Antonia Romeo, tipped to be the Home Office permanent secretary, who was cleared of bullying allegations nine years ago but remains a divisive figure in the civil service. Some of her defenders have said the criticism of Romeo is based on sexist double-standards.
Starmer suggested he wanted to appoint more women to senior leadership roles in government and said he was “determined to transform the culture of government: to challenge the structures that still marginalise women’s voices”.
“And it’s why I believe that simply counting how many women hold senior roles is not enough,” he said. “What matters is whether their views carry weight and lead to change.”
Disclaimer : This story is auto aggregated by a computer programme and has not been created or edited by DOWNTHENEWS. Publisher: theguardian.com
