TRYING TO STOP WOMEN OBJECTIFICATION ON AI COMMUNITY ‘ART’ PLATFORMS IS GOING TO BE A VERY LONG HAUL … (first published in ‘Medium’)

back to 'thoughts'

“This miry slough is such a place as cannot be mended. It is the descent whither the scum and filth that attends conviction for sin doth continually run.” (The Pilgrim’s Progress, 1678)

One of the most challenging aspects of trying to kickstart a mini campaign on an issue most people prefer not to think about is navigating that “slough of despair and despond” so graphically described in John Bunyan’s 1678 Christian allegory, The Pilgrim’s Progress. Now 12 days into my self-imposed crusade, I’m only just beginning to understand the magnitude of the problem. Which is the increasingly widespread abuse of AI text-to-image community ‘art’ platforms to objectify provocatively posing women with exaggerated body parts. A search for images of ‘girls’ in DeviantArt’s online ‘gallery’ is enough to discourage anyone with genuine concerns about this scourge.

And nobody seems to care! As things now stand, AI community ‘art’ platforms are completely unregulated. So, where standards exist at all they seem to be applied inconsistently when it comes to images of women. Adult/not-safe/suitable-for-work (NSFW) portrayals of the female form are bandied about with impunity — tending to suggest that having a set of community standards is merely a box-ticking exercise.

Yet even Dalle-e (a major AI ‘art’ software developer) believes that users calling for unrestricted image generation need to consider “the long-term implications of fully unregulated AI tools”. These are summarised in a Dalle-e-Generator article among other things drawing attention to the importance of:

• preventing harmful image-generation (noting that “AI tools often block explicit, violent, or illegal content”)

• avoiding copyright infringement (noting that “many platforms restrict the generation of copyrighted characters, logos, and artworks”)

• ethical concerns (noting that “AI can be misused to create misleading or deepfake images”), and

• fairness (noting that “restrictions help prevent AI from generating discriminatory or biased content”).

Clearly, online ‘galleries’ don’t give a jot! Featuring an avalanche of images portraying suggestively posing women with wildly exaggerated body parts, they’re accessible to anyone with basic internet browsing skills — including children.

The European Union’s (EU’s) AI Act is often referred to as “the world’s first comprehensive horizontal law governing artificial intelligence” (AIPRM). Partially in force since August 2024, its “staggered” implementation could take another two years. And eventually, the Act is expected to impact significantly on companies conducting business with EU member states.

But following engagements with the US administration, “adjustments” to some of its provisions and “other digital regulations” are now being considered “as part of a wider simplification process” (Reuters) — pointing to the distinct likelihood of enforcement delays and diluted content.

Not that legislating against women objectification will ever be taken as seriously as intellectual property and copyright anyway.

No wonder I’m feeling bogged down …