The harmful impact of generative AI on women

Photo by cottonbros at Pexels

Araoluwa Aina explores how AI is accelerating online abuse and reinforcing gender bias

The rise of Artificial Intelligence (AI), particularly generative AI, in recent years is changing how we interact with each other, how societies function, how technology operates, how we create content, how we approach work and more.

It seems that generative AI is reshaping the world as we know it, to put it lightly. But while we are constantly being told about all the positives, it’s important to consider the possible harmful effects of the AI boom, especially its disproportionate impact on women and girls. One key danger is its ability to accelerate technology-facilitated violence against them.

Even before the recent rise of generative AI, women and girls were already disproportionately affected by online abuse and harassment, including both sexual and non-sexual forms of abuse. These threats are becoming even more prevalent, particularly through image-based abuse. Deepfakes, which are digitally created or manipulated images and videos made to look real, are a clear example. They are increasingly being used to create sexually explicit, non-consensual images and videos of women.

According to a recent UN Women report, 98% of deepfake videos online were pornographic and 99% of the people targeted by deepfake pornography were women. These figures show just how heavily women are targeted by this form of digital sexual abuse. With AI, creating and spreading this kind of abuse has become easier than ever.

We need to think about how AI could be training our minds to normalise gender bias

There are now, tools specifically designed to ‘undress’ women digitally, further intensifying the mass production of deepfake sexual images. As a young woman, it’s unsettling to think that technology can be used to manipulate images of women without their consent, especially when those images can be shared so easily and have such lasting consequences. These images, once created, can be extremely difficult to remove and can have long-lasting psychological and reputational effects on victims.

Apart from the issue of image-based abuse being facilitated and accelerated by AI, there is also a deeper concern about AI itself and the gender bias integrated within generative AI systems. Research by the Berkeley Haas Center for Equity, Gender and Leadership analysed around 133 publicly documented AI systems across different industries and found that 44.2% demonstrated gender bias, while 25.7% showed both gender and racial bias.

This bias is often rooted in the datasets these systems are trained on, which can reflect existing inequalities and stereotypes. Tools used in recruitment are more likely to recommend men over women, especially for higher-paying jobs. In healthcare, these systems can also under-emphasise women’s needs and health issues, contributing to gender-based disparities in care. As we move into the ‘next generation of technology’, with these systems becoming increasingly involved in our daily lives, we need to think about how they could be training our minds to normalise gender bias and intensifying the misogyny that already exists in society.

As AI continues to shape our future, we need to make sure it helps create a more equal society, rather than reinforcing the inequalities we already face.