Mary McGill: Laws on AI sexual abuse images are too late for many victims

Regulations are coming too late for the many people already harmed by AI sexual abuse technology, but a proactive approach could prevent future damage, writes Mary McGill 

The EU ban comes five months after the Grok controversy brought into sharp focus the grim capabilities of so-called ‘nudification’ technologies, apps and websites that use AI to ‘undress’ unsuspecting targets.

On May 7, the European Parliament and Council announced an EU-wide ban on artificial intelligence (AI) systems that "create child sexual abuse material or depict the intimate parts of an identifiable person, or them engaged in sexually explicit activities, without that person’s consent". 

Covering images, audio and video, the new ban will deny access to the EU market to AI technologies with the capacity to create such content. When it comes into effect in December, the ban will also make it illegal to release content-creation technologies "without reasonable safety measures in place to pre-empt the creation of abusive material".

From the perspective of someone researching gender-based online abuse, the ban is a welcome development. However, like many regulatory responses to the worst excesses of new technologies, it comes after the fact, when harms that were predictable have already occurred. 

This offers cold comfort to those hurt, but it does highlight how proactive approaches could potentially reduce future harms.

The EU ban comes five months after the Grok controversy brought into sharp focus the grim capabilities of so-called ‘nudification’ technologies, apps and websites that use AI to ‘undress’ unsuspecting targets.

The mainstreaming of nudification

In late 2025, using prompts like "put her in bikini" users on X (formerly known as Twitter) began to use Grok, an AI chatbot integrated into the platform, to "undress" women and alleged minors, producing an avalanche of gratuitous, nonconsensual sexual imagery.

Nudification technology existed before Grok but embedding it in a mainstream platform turbo-charged its capacity for harm. Compared to pre-existing technology, Grok could produce nonconsensual content at an unprecedented scale, with some estimates putting this ability at over 6,000 images per hour versus less than one hundred for other nudification tools.

Pouring additional fuel on the fire, X is not a fringe app or website; it boasts a vast network of 550 million-plus members and a retweet button, making nonconsensual imagery a lot easier to share on X than elsewhere.

The sheer scale of the image-based abuse occurring on X in early 2026 drew a unified response from policymakers and the public in Ireland, across Europe and globally. Almost everyone seemed to agree that what was happening with Grok was unacceptable and that something needed to be done.

The European headquarters in Dublin of social media company X. Grok has faced backlash for generating sexually explicit images, including non-consensual and child-related content. Photo: Sasko Lazarov / © RollingNews.ie

However, as the days and weeks passed, it became increasingly clear that while a swift response was desired, realising it was not as straightforward as many had hoped. X eventually introduced guardrails to curtail users’ ability to create nonconsensual intimate content but the question of how to effectively address and remedy what happened remains.

The ban on nudification technologies will go some way to help with this. But the simple fact persists that if nudification technologies had, in their early iterations, been prohibited, it would have reduced the likelihood of the functionality becoming embedded in one of the world’s most popular social media platforms.

Researchers had already highlighted the threat posed by nudification technology, not least that the content these apps are instructed to create focuses overwhelmingly on women and girls, is typically deeply objectifying and sexualised, and pays scant attention to consent.

There is also the hard, sad truth that well before the Grok debacle, nudification technologies were already being used to humiliate women and girls. It was only when the abuse occurred at scale on the ‘main street’ of the internet that the urgent need to address it became clear.

Slow-to-act regulators

Regulators' slowness in adopting pre-emptive rather than reactive responses to gender-based abuses online and other forms of online harms is not a new phenomenon. 

The advent of social media and other ‘disruptive’ digital technologies has shown how difficult it can be to rein in companies that scale rapidly across the globe, often with little regard for norms like privacy.

These companies also represent a new vanguard of immense economic, social and political power, adding another layer of complexity for regulators to navigate. 

Mary McGill: 'It was only when the abuse occurred at scale on the ‘main street’ of the internet that the urgent need to address it became clear.'

Nevertheless, as the race to achieve AI supremacy intensifies, it is vital that regulators try to do so, acting proactively wherever possible, before predictable harms occur.

The risks posed by AI are often framed as ‘emerging’. The reality is that the existence of technology like nudification apps is harming women and girls right now. 

We also know the damage social media can cause growing minds, the ways in which algorithms can manipulate our sense of the world and how surveillance capitalism threatens privacy. 

Harms like these shouldn’t need to reach crisis point before regulators intervene. By that point, for too many it is too late.

What's your view on this issue?

You can tell us here

Grappling with these issues is not easy. We are living through the greatest revolution in human communications since the Gutenberg Press cranked up in the 15th century. 

Some lessons will unfortunately have to be learned in hindsight, but where pre-emptive steps can be taken to reduce harms, we must endeavour to do so.

• Mary McGill is a Research Ireland postdoctoral fellow at the Centre for Global Women's Studies, University of Galway