AI ART, FEMALE OBJECTIFICATION & GENDER-BASED VIOLENCE (first published on ‘Medium’)

back to 'thoughts'

could AI text-to-image art be fanning the flames of violence against women?

You may be wondering why, as a South African, I’m trying to raise awareness of the perils of AI text-to-image community ‘art’ platforms here. After all, as an American-owned outlet for social journalism, ‘Medium’ is probably read mostly by Americans. Why not write for a South African blog?

Well, for starters, with a population of roughly 340m the US offers more readership potential than South Africa with its approximately 60m residents — at least 55% of whom are poor (World Bank). On average, a poverty-stricken South African keeps body and soul together on an income of between US$45 and US$86 per month (StatsSA).

By comparison, the US poverty rate is estimated to be around 11% (PovertyUSA), with the income of each indigent American estimated at approximately $1300 per month (HealthInsurance). Of course, the US cost of living is way higher than South Africa’s — but in the context of access to AI art generators, that’s another topic altogether.

In 2023, North America (theoretically the US, Bermuda, Canada, Greenland and St Pierre/Miquelon) held 38% of the global AI image generator market, (Vision). Admittedly, this may well have changed during the past two years, mainly because in 2023 the Asia-Pacific market share was expected to grow at the highest rate. Nevertheless, given the ever-increasing pace of US tech innovation and the fact that most major AI companies are located in the US, it’s probably fair to say that most users of AI art generators and most members of AI community art platforms are American.

With all that in mind, it makes sense to try alerting Americans to the scourge of female objectification among the users of AI text-to-image community art platforms. They have more influence on market trends and, by implication, more power to stop this potentially dangerous scourge.

So, let’s get down to brass tacks. As I wrote in an earlier article, there’s widespread concern that AI text-to-image art is being used to perpetuate female objectification. And being objectifed by others often triggers self-objectification — one manifestation of which is understood to be dressing to attract attention.

“Dress is a conscious act of appearance management”, with clothing style used to “cover or reveal the wearer’s body shape” (Yingqi Wang, Cong Meng and Xiaofeng Jiang, International Journal of Engineering Research & Technology). And in the context of male attention to a woman’s body, dressing provocatively to attract the male objectifying gaze is often a sign of self-objectification — both of which are linked to violence against women (Frontiers).

Interestingly, however, the targets of violence against women may not necessarily be flashy dressers. Some research even suggests that — because sexual harassment and violence against women are “about power” — a provocatively dressed woman exuding “confidence and assertiveness” probably isn’t a would-be harasser’s “ideal target”.

Whatever the case, female objectification has made all women potential targets, regardless of how they dress. Which is because “sexual violence is a consequence of a dehumanised perception of female bodies” acquired through the aggressor’s “exposure and interpretation of objectified body images” (Frontiers).