The App Allows Men to Rate AI-Generated Models, Sparking Sexism Concerns
A Return to Mark Zuckerberg’s FaceMash with a Twist
Please consider spending a minimum of 30 seconds on this page.
This small gesture will significantly support my work.Two decades ago, Mark Zuckerberg, a college student, designed FaceMash — a controversial “hot-or-not” website that became the foundation stone for Facebook.
Fast forward to today, and a developer has replicated that idea. However, there’s a twist: this time, the women are AI-generated.
Welcome to SmashOrPass.ai
To say that smashorpass.ai feels like a déjà vu would be an understatement. A blast from the past, but in the eeriest way. It seems to mock Zuckerberg’s original site, parading itself as an ‘innovative’ experiment with AI image generation. Users are shown images of “women” and are asked for a thumbs-up or thumbs-down.
The catch? These women aren’t real; they’re AI’s creations. But, that’s not the most disturbing part.
Even more disturbing was, the majority of the female representations on the website exhibit exaggeratedly large breasts.
Their faces possess an uncanny, overly polished look that’s characteristic of AI outputs. Distinct outlines of their bodies against the backgrounds further indicate AI’s hand in their creation.
What’s even more unsettling is the depiction of women without faces, just featuring disproportionately large breasts.
Emmet Halm, the man behind this project, tweeted it as a “generative AI party game” and, quite interestingly, left it at that.
It seems he assumed everyone would “get it”. The tweet received significant traction, suggesting many did.
However, this “game” didn’t amuse everyone, especially AI professionals. Sasha Luccioni, an AI expert, expressed her concerns, questioning if technology had really advanced or just looped back to its questionable past.
Another developer, Rona Wang, retaliated with her version, designed to rate men. But not on looks; on their likelihood to be harmful towards women.
The Controversy
This isn’t the first time AI has shown potential biases. These biases, be it racist or sexist, have been thoroughly discussed and criticized.
Yet, it hasn’t stopped developers from launching apps that, unfortunately, reflect these stereotypes.
Some even argue that AI’s being too “politically correct”, referring to tools like ChatGPT’s refusal to use racial slurs.
Halm’s tall claims about his project’s “revolutionary” nature also seem dubious.
He describes it as a “recursively self-improving” engine, adapting to user clicks.
Yet, users have reported repeated images — a sign that the AI isn’t learning. Halm promises this feature in a future update, but for now, it remains absent.
One Twitter user responded,
“Bro wtf is this. The concept of finetuning your aesthetic GenAI image tool is cool but you definitely could have done it with literally any other category to prove the concept, like food, interior design, landscapes, etc.”
While Halm’s tweet suggests he’s proud of trying out “new ideas”, Luccioni and many others believe some ideas are better left unexplored. Why, she questions, when there’s a plethora of neutral data like cars and kittens, did the tool focus on women with exaggerated features?
As advancements in AI continue, one can only hope that future innovations consider ethical implications as much as they do technological feats.
But for now, SmashOrPass.ai stands as a stark reminder that while technology evolves, some of its age-old issues persist.
🔍 Struggling to Find Perfect Medium Publications for Your Article?
