From scissors and glue to code and clicks - AI and sexual violence at work

Over the last two weeks, we’ve seen yet another incident evolve where technology has outpaced accountability and legislation.

GROK, the generative AI chatbot developed by xAI and embedded into X, has once again demonstrated how quickly AI systems can reproduce, amplify and legitimise misogyny, sexualised violence and degradation, often under the banner of “free speech”, humour, or provocation. Much of the public commentary has focused on whether GROK is politically dangerous, whether it spreads misinformation, or whether it embarrasses its creator.

The real harm of generative AI is not primarily political destabilisation. It is sexual violence, disproportionately directed at women and girls, delivered through new tools that feel novel but are rooted in very old power dynamics.

As someone who works every day with organisations responding to sexual misconduct, harassment and abuse, I can say this clearly: AI has not changed why harm happens. It has changed how easy it is to perpetrate it, how far it spreads, and how devastating the impact can be.

From scissors and glue to code and clicks

I often use two examples in training because they expose the continuity between “old” and “new” harms.

In one scenario, a man physically cuts out the faces of his female colleagues and sticks them onto nude bodies taken from magazines. He pins them to a staff noticeboard. Colleagues laugh, place bets, rate the women. It is degrading, humiliating, hostile.

In another, a colleague uploads a woman’s photo into a free AI app. Within seconds, a pornographic deepfake is generated. It is shared in a WhatsApp group. People mock her. Screenshots circulate. The impact follows her far beyond the workplace.

One feels archaic. The other feels modern. They are both real and have happened. Fundamentally, they both sit on the same continuum of harm.

Both acts are about entitlement to women’s bodies. Both are about humiliation as control. Both rely on an audience. Both escalate because they are normalised.

Technology did not create the intent. It removed friction.

Normalisation is how violence escalates

Sexual violence rarely begins with what we recognise as “serious”. It begins with jokes. Comments. Banter. A culture that tolerates small acts of degradation because “no one seemed bothered”.

Over time, the behaviour escalates. What once felt shocking becomes familiar. Silence becomes misread as consent. And because nothing bad happened last time, perpetrators feel emboldened to go further.

This is how we end up saying “How did this happen?” Because we only notice when the harm becomes undeniable.

AI accelerates this process. What once required time, effort, access or courage can now be done in seconds, anonymously, at scale. The law is scrambling to catch up because the harm is already here.

In the UK, the fast-tracking of legislation to criminalise the creation of sexually explicit deepfakes without consent is a recognition of just how severe this problem has become and employers must take note.

What young people are telling us that we are not hearing

One of the most striking shifts I’ve seen in recent years comes from 18–25-year-olds in training sessions, particularly apprentices and early-career employees.

They talk about deepfakes casually. About AI-generated nudes. About sexual humiliation in group chats.

Not because it doesn’t affect them, but because it is so common and real for them.

Many have already experienced it before entering the workforce. They don’t always name it as abuse. They don’t always see it as criminal. It is just “how things are”.

That should terrify employers.

Because when people do not recognise themselves as victims of sexual crime, they are far less likely to report. And when organisations fail to recognise digital sexual harm as serious, illegal and deeply traumatic, they inadvertently reinforce the very cultures that allow it to flourish.

GROK, Elon Musk and the myth of neutrality

It is tempting to talk about GROK as a glitch, a misstep, a teething problem of innovation. That is generous.

AI systems reflect the values they are built within and the constraints they are given. When “edginess”, provocation and engagement are rewarded, the output will follow. When accountability is framed as censorship, harm will be reframed as humour.

This is not about demonising technology. It is about refusing to pretend that technology is neutral when it is deployed without ethical restraint.

And it is about naming what is happening honestly: sexualised degradation of women is being automated, scaled and normalised under the guise of innovation.

Why this matters for the workplace now

The Worker Protection Act has changed the legal landscape. Employers now have a proactive duty to take reasonable steps to prevent sexual harassment.

That duty cannot be met by policies designed for a world where harm only happened in corridors, at desks or at Christmas parties.

Sexual misconduct now happens:

• on phones

• in group chats

• through AI tools

• outside working hours

• without physical proximity

And the impact does not stop at the screen.

If organisations treat this as “online behaviour” that isn't their problem rather than sexual violence that has a possibility for perpetration connected to the workplace, they will fail both their people and the law.

This is not about banning technology. It is about understanding perpetration. Sexual violence is about power and control. AI simply offers new ways to exert it.

Whether the tool is scissors or software, the intent remains the same. And intent, not technology, is where accountability must begin.

Interested in this topic? Click to read more.