Louisa Street, School of Law, Keele University, United Kingdom, louisa.a.street@outlook.com
Louisa Street is a youth worker and PhD candidate with over 15 years’ experience working with young people on a range of issues including sex and relationships, drugs and alcohol awareness and online safety.
New Year, new digital controversy with Elon Musk at the centre. Since the start of 2026, there has been uproar about Grok, X’s AI, creating sexualised images of women and children. This has led to debates about whether current laws are sufficient to regulate this technology. As a PhD student exploring the law around the digital exchange of sexual images, I’ve been interested to see politicians stating they will make new, tougher laws to regulate AI, rather than enforcing the existing legislation. And it leaves me with a bigger question unanswered: what can we do about the attitudes that make so many people think this behaviour is acceptable?
Before I explore a potential response to this question, I’ll outline the current legal situation.
Indecent Images of Children
The Protection of Children Act, 1978 made it an offence to create, distribute or possess an indecent image of a child. Whilst the original offence stated that a child was anyone under 16, in 2003 this was amended to be anyone under 18. When this amendment was made, a caveat was added, that if the person in the image was married to, or in an ‘enduring relationship’ with the person who created or possessed the images, no crime had taken place.
In 1995 the reference to ‘pseudo’ images was added, meaning any indecent image, whether or not it looked, or indeed was, real, was covered by the legislation. In 2003 the notion of ‘possessing’ such an image was broadened to include deliberately viewing an image on a digital device, even if that image was not downloaded or saved. In addition, the Serious Crime Act, 2015 s.69 made it an offence to share information about how to create such images, which would include sharing advice on prompts for AI to generate images.
The Online Safety Act 2023 s10 (2) adds an additional point to this, and that is that for the first time, platforms will be held legally responsible for removing any of the above mentioned illegal content before it is seen by users. At s10 (5) the Online Safety Act also makes platforms legally responsible for removing information about how to create or obtain indecent images of children. Prior to the Online Safety Act, platforms removed illegal content in line with their own community guidelines, sometimes referred to as self-regulation.
To protect teenagers who consensually share images of themselves with a peer from undue criminalisation, in 2016 the police developed Outcome 21. This allows police to close a case which would otherwise lead to a charge under the Protection of Children Act, provided images were not distributed more widely. It was developed on the grounds that it would not be in the public interest to charge young people as sex-offenders for exploring their sexuality through taking and sharing images of themselves with peers, which can be part of typical healthy sexual development.
Non-consensual Intimate Images (NCII) of Adults
The Online Safety Act, 2023 s.188 added a new offence to the Sexual Offences Act, 2003, s.66B amending existing attempts to legislate against so-called ‘revenge porn’, more accurately described as non-consensual intimate image distribution. This law makes it an offence to share an intimate image of another person without that person’s consent. Previous legislation had specified that this was an offence only if the person distributing the image intended to cause distress to the person pictured. Campaigners argued the intent of the perpetrator is irrelevant to the harm experienced by the victim, which led to the amendment in the Online Safety Act. The Data (Use and Access) Act, 2025 s.138 criminalises the creation of digital images without consent, but has not yet come into force (although MPs have promised that these will be brought into force imminently).
A New Law for AI?
Some have argued that there should be more legislation to prevent people from creating illegal material using AI. There are potential benefits of to such legislation, and the argument that AI has run ahead of our ability to regulate it are compelling. But new legislation alone won’t prevent the creation of such images: all those who used Grok to create sexual images of under eighteens could be prosecuted for creating, possessing and distributing offences under the Protection of Children Act, 1978. Anyone who deliberately viewed indecent images of under eighteens could be prosecuted for the possession offence under the Protection of Children Act, 1978. Anyone who created and then shared a sexual image of an adult without consent could be prosecuted for an offence under the Sexual Offences Act, 2003. And Grok / X could be prosecuted for not removing this illegal content before it was seen under the Online Safety Act, 2025.
Given that this suggests there are already ample legislative options to pursue those who created sexual images using AI, the discussion of the need for more legislation may be a case of politicians wishing to seem ‘tough on crime.’ But if legislation alone isn’t the answer, what is? The new 20 mph speed limits in residential areas may offer a useful analogy. Everyone who drives a car has had to take a test to prove they understand the rules of the road. So, everyone driving a car knows that it is illegal to exceed the speed limit. However, occasionally you may be in a rush, you may not notice that there is a new speed limit, or you might just feel like you should go faster than 20 mph. Evidence from a Parliamentary briefing suggests that 20 mph speed limits reduce casualties, whilst also encouraging people to walk and cycle more, tangible benefits that we should be pursuing. But the same briefing states that more than just signage is needed to see real compliance with these speed limits, including changes to physical environment (such as speed bumps) education campaigns and enforcement.
A recent example of a ‘speed bump’ in the digital world was the experimental warnings and links to support for those searching for illegal pornography. If a user on Pornhub searched for certain terms linked to non-consensual images, or those of children, a warning was displayed with a link to support services. The pilot project found that 82% of those who were shown such warnings once never triggered another warning, and instead searched for other, legal content. They also note that over 1500 people sought information, advice or support from the Lucy Faithful Foundation (who offer support to potential perpetrators to prevent child sexual abuse) after seeing the warning.
Similar safeguards could be built into AI systems to prevent users creating illegal sexual images. However, these safeguards alone will not solve the problem. We need to explore the culture that enables this behaviour – the demand, not just the supply. That demand is based on a culture where people feel it is acceptable to make nude images of others (overwhelmingly women and girls). A culture in which people feel entiltled to the see the bodies of, regardless of whether that person consents to be seen that way. Blocking supply through mainstream platforms without reducing demand, may push people to less mainstream platforms.
A Safe Uncertainty and Children’s Rights Solution
For most people, especially professionals supporting young people, creating legislation for AI systems is not within our power. And even if it were, viewing legislation as the answer would be an example of attempting to utilise a ‘safe certainty’ approach, where we tell young people ‘this behaviour is illegal’ rather than exploring the issues that lead to images being created and shared non-consensually. Unfortunately, attempting to remove all risk, without creating space to explore the social and ethical influences underpinning the behaviour, despite our best intentions, inevitably leads to ‘unsafe certainty,’ where young people are left without the tools to navigate the underlying issues.
Taking the safe uncertainty approach requires conversations and openness to different viewpoints. This is not to say we accept or condone the creation of illegal images. Instead, we need to recognise that young people may face issues within their peer groups linked to Image Based Sexual Abuse (IBSA) and need our support to navigate those issues and the potential conflict that may arise from it. We need to be open to discussing things which make us uncomfortable, and model ways to do it that enable young people to respond if they are faced with peers who may claim, for example, that using AI to create sexual images is a bit of harmless fun – they are not real, so are somehow less harmful. Refusing to allow this viewpoint to be expressed in the classroom may feel safe for us but leaves young people with no ability to deal with it in the real world.
In the training I co-deliver as part of the Headstart Kernow digital resilience project, I’ve developed the an approach which uses the UN Convention on the Rights of the Child (CRC) as a framework for discussing sexual image sharing with young people.
Applying this to the context of AI generated intimate images, I would suggest this involves:
Inform: Telling young people about the law, but also about the support services that are available, including Shore Space, for those who are worried they might commit an offence as well as intimate image removal services like Take It Down for under eighteens, and Stop NCII for adults. These allow victims of IBSA to report the abuse and have their images removed from public sites (like X). Teaching young people about their rights and reassuring them that we won’t be criminalising them unnecessarily.
Challenge: Tackling discrimination whenever you see it, as those who are discriminated against due to race, sexuality or (dis)ability are more likely to have their intimate images non-consensually shared. Discussing with young people that they have a right to express themselves, and that includes sexual expression, but they also have a right to keep that private (and have their privacy respected by others).
Focus on Consent: Whilst young people may not be legally able to consent to sharing sexual images until they are 18, the CRC gives them the right to have a say in matters which affect them, including in matters of expression and privacy. Added to this, the Fraser Guidelines suggests consent is not simply bestowed on one’s eighteenth birthday, but a skill which develops throughout adolescence. Finally, the CRC says we must teach young people to respect the human rights of others, that includes respecting other people’s right to consent or withhold consent.
Whilst politicians and campaigners may say we need more, better, or tougher legislation to prevent the non-consensual creation and sharing of intimate images, this blog has outlined the wealth of legislation that already exists, suggesting that more legislation alone will not solve the problem. Instead, we need to start challenging the notion that non-consensual intimate image distribution is inevitable, and only the threat of prosecution will stop people doing it. Using a rights-based approach, we can inform young people of their rights, we can challenge behaviours which violate these rights, and we can teach them to view consent as theirs to give or withhold. We can teach them that everyone has the right to consent (or withhold consent). And that people’s right to consent should be respected, rather than being something that can be overcome, overpowered or coerced.