What gets flagged? The same thing that’s always been policed: visibility outside the script.
A breastfeeding photo is removed for “nudity,” but a lingerie ad with more skin stays up.
A figure-drawing post is buried; a “try-on haul” is monetized. Same anatomy. Different framing.
Why? Because moderation is a stack of incentives, not a moral compass.
- Legal risk beats educational value.
- Brand safety beats nuance.
- Automation beats context.
Machine learning doesn’t understand consent, intent, or culture. It scans for surfaces—skin tone, contours, contrast—then guesses. And when it guesses wrong, who pays? Small creators, educators, independent artists. People without a hotline to an ad rep.
Meanwhile, violence slides through because it’s legible to the system: action, conflict, shareable outrage. A body resting in its own presence is harder to categorize—and easier to punish.
When we say “the algorithm did it,” we let humans off the hook. Policies chose the thresholds. Leaders chose the objectives. We could tune models to distinguish art from exploitation, education from erotica, presence from performance. But that takes effort—and admitting the current line isn’t about safety. It’s about liability.
Here’s the gut check: if a platform consistently hides non” sexual bodies” while spotlighting harmful spectacle, it’s not protecting the public. It’s protecting its business model.
And if that sounds harsh, good. Because until we say it plainly, nothing changes.