GOING FORWARD: spreading the word ... (networking for likeminded people)

back to 'thoughts'

speaking up and out ... don’t allow yourself to be manipulated into believing what’s popular is necessarily OK

Anyone interested in taking this mini-campaign to dizzier heights may want to contact Julia Reinhardt, a San Francisco and Berlin-based consultant specialising in AI governance and public policy. Her website has a message facility, which I've just used. I wonder if she’ll reply? Watch this space! (she did – but only to say that the article I found helpful was written by another Julia; a few days later, she suggested I contact Alexandra Wudel at Fem-AI)

Hi, Julia (9 November 2025)

I’m concerned about the proliferation of images published on the public galleries of international AI text-to-image community ‘art’ platforms objectifying women with exaggerated body parts in suggestive poses. As a member of one such platform, I have tried to raise awareness internally of the perils of this – to no avail. I have extensive correspondence with platform management and moderators about vague community standards and widespread non-compliance. On principle, I do not intend renewing my subscription.

That said, these platforms appear not to be regulated at all. Also, my sense is that the EU AI Act may not have made adequate provision for measures to mitigate women objectification and related social ills. And if the EU Digital Services Act applies to international community platforms, the one to which I belong may not be compliant. Yet it has European members (mostly German, French and Swiss).

I’m South African and have begun trying to raise awareness of this in my own small way using what I hope is an eye-catching website with a blog, along with articles in Medium and posts on X. Because the platform to which I belong is Australian, I have also alerted their Competition & Consumer Commission and Law Reform Commission.

An especially disturbing American platform (DeviantArt) seems to be setting a trend now being followed by the Australian platform and likely to be mirrored worldwide. This despite a plethora of empirically based research on the psychological and other damage done by women objectification.

I read your article in Picsagon, (apparently written by another Julia, as I was told in no uncertain terms by email) which is what prompted me to visit your website and make contact. Would you be willing to provide the names of influential lobbying organisations concerned about the same issues?

Sincerely,

Pam Saxby

Hi, Fem-AI (18 November 2025)

Julia Reinhardt suggested I make contact, although I don’t know her personally. I found her website during some desk research on women objectification in AI ‘art’. When I asked for pointers to people with enough influence to stop it, your name came up.

As I’m sure you know, there are lots of online community platforms offering subscribers access to a range of AI ‘art’ generating models. DeviantArt (based in the US) is one that’s especially crass about publishing subscriber-created images of oversexualised women. Another is Australian-based NightCafé. As far as I can tell, these platforms are completely unregulated.

NightCafé (to which I subscribe) relies on auto-moderating filters to spot ‘not safe/suitable for work’ (NSFW) adult content – with human moderators stepping in to catch images that fall through the cracks and are flagged by community members post-publication. The problem is that – although community standards require subscribers to tag suggestive images of women with exaggerated body parts ‘NSFW’ (in which case they aren’t made available for general viewing) – NightCafé tends not to apply its own standards consistently.

This may well be because adult AI imaging is a money-spinner, which would explain the extent to which the platform encourages and even promotes women objectification in many of the games/challenges it runs. Of course, this is an age-old issue. But my sense is that AI is making it far worse – with all kinds of negative implications for gender stereotyping and related social ills.

I have already alerted Australian and US government agencies to my concerns. However, as you know, nothing can be done without legislative measures in place to curb AI image generating related abuse.

Are you aware of any EU interventions in the pipeline? If so, I would appreciate some pointers so that I can alert them to the problem.

Regards,

Pam Saxby

Of course, being fully conversant with the issues at stake helps. And who better to enlighten us than major AI text-to-image ‘art’ software developer Dalle-e?

Apparently, ‘most AI image generators use content moderation filters to prevent the creation of restricted content’ (Dalle-e-Generate). These filters rely on:

• pre-trained datasets (“AI models are trained to recognise and block certain types of images”)

• automated content scanning (“AI scans generated images for violations before allowing downloads”)

• user guidelines and policies (“platforms set clear rules on what can and cannot be generated”).

Great! So, how did we get to this (DeviantArt)? (if you find women objectification offensive, think carefully before searching this ‘gallery’ for ‘girls’).

Well, according to Dalle-e-Generate, “while these measures help maintain ethical AI usage, they also limit creative freedom for some users”.

So, artistic freedom’s the nub of the problem. The developers may well be doing their best to protect women (and children) from AI image-based objectification – but some community ‘art’ platforms aren’t. And some users are taking advantage of that.

I hope my previous blog posts and the information I provide in the pages of this website will motivate other people concerned about woman objectification to lobby for the industry’s more effective regulation. Establishing an international AI community ‘art’ platform ombud might also be worth considering.

Meanwhile, “the debate between unrestricted AI image generation and ethical responsibility will continue to evolve”.

Concerned visitors to this website and readers of its blog may find some comfort in the fact that – as a major AI ‘art’ generator developer – even Dalle-e-Generate believes that users calling for unrestricted image generation “must consider the long-term implications of fully unregulated AI tools”. We touch briefly to some of these here and here and here.

GOOD LUCK!