Sitting at my laptop I watched a naked woman with my face having graphic penetrative sex in a variety of positions with a nude man. The pornographic video lasted for three minutes and 32 seconds, and grotesque as I found it, I made myself watch it all. I needed to understand exactly how realistic these images are, and also recognise how easy it was for people to access them online.
Because as seamless as the footage appeared, it wasn’t me at all – my face had been imposed on to another woman’s body using artificial intelligence (AI) to make what is known as ‘deepfake’ pornography.
The video had been unearthed by my Channel 4 colleagues while researching the exponential and alarming rise of deepfake porn for a special report which was broadcast last month.
A deepfake pornography video of Channel 4 broadcaster Cathy Newman was uncovered by her colleagues while researching the rise of the technology
Out of 4,000 celebrities they found featured in deepfake porn videos online, 250 were British – and one of them was me.
None of the celebrities we approached to comment on this would go public. While disappointing, I understood – they didn’t want to perpetuate the abuse they’d fallen victim to by drawing more attention to it.
But for our investigation to have maximum impact, I knew that I needed to speak out.
In my 18 years as a Channel 4 journalist I have, sadly, seen plenty of distressing footage of sexual violence. So while I was nervous about becoming part of the story, I thought I would be inured to the contents of the video itself.
But actually it left me disturbed and haunted. I’d been violated by a perpetrator whom, as far as I know, I’ve never met, and I was a victim of a very modern crime that risks having a corrosive effect on generations of women to come.
I also felt vindicated by my decision to go public, because earlier this month the Government announced that the creation of these sexually explicit deepfakes is to be made a criminal offence in England and Wales.
I understand that Laura Farris, the Minister for Victims and Safeguarding, was motivated in part to take action after watching our investigation. This comes after the sharing of this type of content was outlawed in the Online Safety Bill last year.
My colleagues were already researching deepfake pornography when, in January, fake explicit images of the singer Taylor Swift went viral on X/Twitter, with one image viewed 47 million times before it was taken down.
Suddenly the alarming scale of the problem became clear. We found the four most popular deepfake porn sites hosting manipulated images and videos of celebrities had had almost 100 million views over just three months, with more deepfake porn videos created in 2023 than all the years combined since 2017.
The videos have been viewed in total more than 4.2 billion times.
You might think some degree of technical expertise is required to make them, but it’s incredibly easy and done mostly using smartphone ‘nudify’ apps – there are more than 200 available. Users submit a picture – one single photograph of someone’s face grabbed from social media is all that’s needed – and this is used to create a horrifyingly realistic explicit image.
Because of the sheer number of celebrity pictures online, we hear about high-profile personalities becoming victims most often. They include American congresswoman Alexandria Ocasio-Cortez, who this month described the trauma of discovering she had been targeted while in a meeting with aides in February, and Italian prime minister Giorgia Meloni, who is seeking damages after deepfake videos of her were uploaded online.
But arguably the greater victims are the hundreds of thousands of women without a public platform to denounce the images as deepfake – the women who might be in a meeting or job interview and not know whether the people opposite them have seen and been taken in by the fake footage.
The recreation of the broadcaster. Out of 4,000 celebrities they found featured in deepfake porn videos online, 250 were British – and one of them was me, Cathy writes
I spoke to one such victim, Sophie Parrish, 31, a florist and mother-of-two from Merseyside, whose deepfake porn video was uploaded to a website by someone close to her family, which men then photographed themselves masturbating over. She was physically sick when she found out, and the impact on her since has been profound.
A beautiful woman, she’s lost confidence and now doesn’t want to put on make-up for fear of attracting attention. She almost blames herself, although obviously there’s no blame attached. And yet she had the guts to go public last February, petitioning the Ministry of Justice to make it illegal to create and share explicit images without consent.
In truth, I wasn’t entirely surprised when my colleagues told me about the existence of my video, given that, as a woman in the public eye, I have been trolled relentlessly for years.
After my interview with Jordan Peterson, the Canadian psychologist infamous for his divisive views on political correctness, free speech, gender identity and racial privilege, went viral in 2018, I received death threats. I was called a ‘c***,’ ‘b****’ and ‘w****’ and my eldest daughter, then 13, was distressed to come across a meme on Instagram in which my head had been imposed on a pornographic image.
So it is understandable that my colleagues were keen I didn’t feel under any pressure to watch the video that had been made of me, while my editor was concerned about its emotional impact. But I felt I owed it to each victim of this crime – especially Sophie Parrish, who I’d interviewed the day before – to understand for myself how it felt to be targeted, and speak out.
Of course, I have access to professionals who can help me process the material, but many women – and 98 per cent of deepfake pornography victims are female – do not. I was worried about the backlash on my daughters, now aged 19 and 15, but like all teenagers they’re aware of the kind of AI content proliferated online and were interested as to how we can navigate it.
After watching the report, they told me they were proud. So too was my husband – although he understandably didn’t want to watch the unedited video of me, and nor did I want him to.
While the pornographic meme that my daughter saw in 2018 was crude, I discovered that, six years on, the digital terrain has changed and the boundaries between what’s real and what isn’t have blurred.
The one saving grace from my surprisingly sophisticated deepfake video was that AI can’t – yet – replicate my curly hair, and the bleached blonde bob clearly wasn’t mine. Nonetheless, the footage of me having sex with a man who, presumably, had not given his consent for his image to be used either, felt incredibly invasive.
But I also wanted to be filmed while watching it, to show in our report the extent of its impact on me.
Although it had obviously been made remotely, by a perpetrator whose motives I can only speculate on, I felt violated.
Anyone who knows me would realise I wouldn’t be involved in making a porn video, and one advantage of getting older is that you’re less troubled by puerile abuse. But its existence undermines and dehumanises women. It is a deliberate attempt to belittle and degrade. Even if they know they’re watching deepfake porn, men don’t seem to care.
Seventy per cent of viewers visit deepfake porn sites via search engines. When we contacted Google, a spokesman said they understood how distressing the images can be, that they’re developing additional safeguards to help people protect themselves, and that victims can have pages that feature this content removed from search results.
Since our investigation, two of the biggest deepfake sites – including the one hosting my video – blocked UK users from accessing their content. But the video is still available using a virtual private network – a VPN – that hides a user’s location.
The Government’s legislation to outlaw the creation of these videos – which will result in a criminal record, fine and potential jail sentence, and will be introduced as an amendment to the Criminal Justice Bill – is groundbreaking, but experts I have spoken to have already warned of potential loopholes.
Victims will have to prove the video was made with an intention to cause distress, which may be difficult, and there is a question mark as to whether, if you ask an app to make the explicit content, you are off the hook in the eyes of the law.
Another drawback is that a lot of these videos are made outside the UK, where our legislation won’t apply, so global action is needed as well.
Then there’s the question of timing: Ofcom, the broadcasting watchdog, is still consulting on the rules of the law that made sharing these videos illegal. It won’t come into force until the end of the year, by which time hundreds of thousands more women will have become victims.
Regulation is also lagging far behind the technology that is enabling this crime, so ultimately it comes down to the big tech companies disseminating this explicit content, which is driving viewers and advertisers to its platforms, for profit.
They are far more powerful than individual jurisdiction, and I don’t see any evidence that they’re tackling the issue with the urgency it requires.
I believe it’s in their power to stop these videos circulating straight away, but it’s not in their interests to do so.
I was worried about the potential backlash to making myself a part of this story, but the overwhelming reaction has been supportive, whether on social media, my email inbox or in the street.
And a month on, as depressed as I am at the corrosive effect of AI on generations of women to come, I’m glad I went public.