Is AI Being Used to Target Women Online?
A new UN report says yes — and most of AI's builders still aren't women.
Key takeaway: Yes — a new UN Women report found that almost one in four incidents of online violence against women human rights defenders, activists, and journalists now involve AI, including deepfakes and the non-consensual sharing of intimate images. At the same time, women hold only 22% of AI jobs worldwide and…
By Dear Sarah · · Updated
Key takeaways
- A 2026 UN Women report found that AI was involved in almost one in four incidents of online violence against women activists, journalists, and human rights defenders.
- Twelve percent of the women surveyed reported non-consensual sharing of intimate images, and six percent said they were targeted with AI-generated deepfakes.
- Women make up only 22% of AI professionals worldwide and hold fewer than 14% of senior AI leadership roles, while men hold close to eight in ten AI jobs.
- AI models trained on male-dominated data sources, like Reddit, tend to produce gender-skewed results because the underlying training data itself is skewed.
- StopNCII.org lets anyone register photos as a digital fingerprint, not the image itself, so participating platforms can block or remove non-consensual copies automatically.
If you've ever gotten that cold, sick feeling after your photo turned up somewhere you never put it — or sat with a friend while she went through it — this one's for you. A new UN Women report, released this month, confirms something a lot of women already suspected: AI isn't only changing how we work. It's changing how we get hurt, and how fast.
The headline finding: almost one in four incidents of online violence against women human rights defenders, activists, and journalists now involve artificial intelligence — think deepfakes, non-consensual intimate images, AI-generated sexual harassment. Meanwhile, women hold less than a sixth of the jobs shaping the technology doing the harming.
Is AI being used to target women online?
Yes, and the numbers are new. UN Women's Gender Snapshot 2026, launched this month at the UN General Assembly, tracked online violence against women activists, journalists, and human rights defenders and found AI involved in roughly one in four incidents. Twelve percent of those women reported the non-consensual sharing of intimate images. Six percent said they'd been targeted with deepfakes. More than one in four had received unsolicited sexual advances through digital messaging, some of it AI-generated.
"The future of AI cannot be built by half the world's talent," said Sarah Hendriks, UN Women's Director of Policy, Programme and Intergovernmental Support, in the report's release. "As artificial intelligence reshapes economies, societies and daily life, women's equal participation is crucial to its design, governance and regulation."
That line lands harder next to the other number in the same report: women make up only 22% of AI professionals worldwide and hold fewer than 14% of senior leadership roles, while men hold close to eight in ten AI jobs. We've written before about why the AI industry still hasn't closed that gap — this report is the clearest evidence yet that it isn't just about pay or promotions. It's about who ends up getting hurt by tools other people built.
Why does AI keep getting this wrong?
Because it's trained on a lopsided slice of the internet, mostly by people who don't experience the harm themselves. Shubhi Rao, CEO of the AI company Uplevyl, explained it simply in a recent Washington Monthly piece: AI models learn from "baskets" of training data, and when one basket is bigger than the other, the output ends up "over-indexing toward one gender over the other, statistically speaking." Roughly 60% of Reddit's users are men, and Reddit is one of the biggest sources of text AI models train on. It's the same root cause behind AI quietly downgrading writing that sounds like how women talk at work. Different symptom, same cause: the input shapes the output, and right now the input is mostly men.
Here's why that's not just an industry story. You don't have to be famous for any of this to reach you. A deepfake or a stolen photo doesn't need you to have a public platform — it needs you to be findable, which most of us already are. Women organizing around causes, like the renewed push to recognize gender apartheid as a crime against humanity at the UN, are now facing AI-assisted versions of the same old intimidation that's tried to push women out of public life for decades. It's just faster now, and more convincing.
One thing to try today
Register your photos at StopNCII.org before you ever need to, not after. It's a free tool from a UK child-safety charity that turns your images into a digital fingerprint directly on your own device — the actual photo never leaves your phone. If someone later tries to upload one of those images to a participating platform, it gets blocked automatically. It takes about five minutes, and you don't have to wait for something to happen to do it.
Quote to sit with
"If you have a face, you have a place in the conversation about AI." — Joy Buolamwini, from Unmasking AI
💌 Sarah
Frequently asked questions
What percentage of online violence against women involves AI?
According to UN Women's Gender Snapshot 2026, almost one in four incidents of online violence against women human rights defenders, activists, and journalists now involve artificial intelligence, including deepfakes and non-consensual image sharing.
Why does AI treat women differently?
AI models learn from training data that skews heavily male — roughly 60% of Reddit's users are men, and Reddit is a major source of AI training text. When one gender dominates the data, the model's outputs reflect that imbalance.
How many women work in AI?
Women make up about 22% of AI professionals globally and hold fewer than 14% of senior leadership roles at AI companies, according to UN Women's 2026 report. Men hold close to eight in ten AI jobs.
What is StopNCII.org and is it safe to use?
StopNCII.org is a free tool run by a UK child-safety charity that creates a digital fingerprint of your photos directly on your device — the images themselves are never uploaded. Participating platforms use that fingerprint to block or remove matching images if someone tries to share them without consent.
Can AI be used to create deepfakes of real people?
Yes. UN Women's 2026 report found that 6% of the women human rights defenders, activists, and journalists surveyed had been targeted with AI-generated deepfakes or manipulated images or video as a form of online harassment.
Sources
Keep reading
- Does AI Make Women Sound Less Professional at Work? — New research shows AI chatbots quietly downgrade writing that sounds like how women actually talk. Here's what to do about it.
- Why Are Women Leaders Saying AI Needs More Humanity? — Nearly 650 women just spent two days at MIT arguing AI is missing something essential. Here's what they mean.
- Does AI Give Women Worse Financial Advice Than Men? — You ask AI for money advice because it feels judgment-free. New research says the answer changes depending on how female you sound.
- Can AI Actually Close the Gender Gap in Women's Health? — Everyone's calling AI the fix for women's health data gaps. Seven researchers just said: not automatically — here's what actually has to change first.