How legitimate sexual health educators are fighting an uphill battle against AI moderators that can’t tell the difference between education and exploitation
Dr. Sarah Chen had been explaining IUD insertion procedures on Instagram for three years when her account suddenly went dark. No warning. No specific violation cited. Just a vague notice about “adult nudity and sexual activity” that left her 45,000 followers—many of them young women seeking reproductive health information—without access to her educational content.
“I was showing a medical diagram,” Chen says, still incredulous six months later. “Not even a photograph—a clinical illustration from a textbook. The same image hanging in every gynecologist’s office in America.”
Chen’s story isn’t unique. Across social media platforms, sexual health educators, medical professionals, and reproductive rights advocates are watching their carefully crafted educational content vanish into the algorithmic void. Recent Associated Press reporting reveals a pattern: legitimate health information is being swept up in platforms’ increasingly aggressive attempts to police sexual content, with fixes only arriving after public pressure or media attention.
The irony is painful. At a time when young people increasingly turn to TikTok and Instagram for health information—with new 2025 research confirming social media as a primary source for sexual and reproductive health knowledge among youth—the platforms are making it harder than ever to find accurate, professional guidance.
The Machines Can’t Tell Medicine from Pornography
Here’s how the system actually works, and why it keeps failing the people who need it most.
Every piece of content you post gets scanned by artificial intelligence trained to spot “sexual” or “nude” content. These AI filters operate on what engineers call “high recall”—they’d rather block a thousand legitimate posts than let one genuinely harmful image through. It’s an understandable priority when child safety is at stake. Meta recently celebrated removing millions of accounts that sexualized minors, and nobody’s arguing against that.
But here’s where it breaks down: these automated systems can’t understand context. They see the word “vaginal” and flag it, whether it’s in a porn title or a medical journal. They spot a breast and sound the alarm, unable to distinguish between sexual content and a breast cancer awareness post. They detect skin tones in certain proportions and configurations, and suddenly your before-and-after surgery photos are “adult content.”
“The AI sees patterns, not meaning,” explains a former content moderator who worked for a major platform and requested anonymity. “It’s looking for keywords, flesh tones, certain shapes. It doesn’t know you’re a doctor. It doesn’t know you’re teaching. It just knows you used words and images that correlate with sexual content in its training data.”
Even when content doesn’t technically violate platform rules, it often gets slapped with what Meta calls “borderline” labels—a shadowy designation that dramatically reduces your reach without notification. Your follower count stays the same, but suddenly only a fraction of them see your posts. You’re broadcasting into an ever-shrinking void, and you might not even know it.
The Global Crackdown Nobody Asked For
The censorship isn’t random—it follows patterns that reveal troubling biases about whose health matters and what information deserves to spread.
Abortion information faces particularly aggressive removal, even in countries where abortion is completely legal. The AP documented cases of posts explaining how to access abortion pills or find clinic services disappearing without explanation, only to be restored after journalists started asking questions. In the Middle East and North Africa, reproductive health content faces even steeper challenges, with legitimate medical information routinely blocked.
Indigenous communities have watched their traditional health practices and body-positive content get flagged as inappropriate. The Oversight Board—Meta’s independent review body—has repeatedly overturned the platform’s decisions, finding that nudity policies were being inappropriately applied to non-sexual Indigenous content.
Gender-affirming care information hits similar walls. Trans creators sharing their surgical journeys or hormone therapy experiences report constant takedowns, appeals, and the exhausting cycle of fighting to keep educational content online. Again, the Oversight Board has had to step in, reversing Meta’s decisions and highlighting how current policies fail marginalized communities.
Medical professionals report a widespread pattern: clinical imagery gets flagged as sexual content, patient education materials disappear, and years of carefully built professional presence can vanish overnight. Plastic surgeons can’t show results. Dermatologists can’t display skin conditions. Gynecologists can’t explain procedures. The platforms’ policies technically allow educational content, but the enforcement reality tells a different story.
Welcome to the Appeals Maze (Population: Frustrated)
So your educational post got removed. What happens when you try to fight back?
First, you’ll likely face an automated review that takes seconds and changes nothing. The same AI that flagged your content initially takes another look and—surprise—reaches the same conclusion. You’ll get a template response that doesn’t address your specific situation or acknowledge the educational context you carefully provided.
If you persist, you might eventually reach a human reviewer. But here’s the thing: these reviewers are processing hundreds of cases per day, often with just seconds to make each decision. They’re following the same conservative guidelines that led to the initial removal. They’re not medical professionals. They might not speak your language fluently. They’re almost certainly not going to spend time researching whether your content has legitimate educational value.
Meta’s own transparency reports acknowledge these “error rates,” tracking how much content gets restored after wrongful removal. But that data doesn’t capture the creators who give up, the time lost, the audiences dispersed, or the critical health information that never reaches the people who need it.
For the lucky few, there’s the Oversight Board—think of it as Meta’s Supreme Court. But only a tiny fraction of cases ever reach this level, and decisions can take months. The Board has consistently sided with health educators, overturning Meta’s removals and calling for policy changes. But individual victories don’t fix systemic problems.
The Underground Railroad for Sex Ed
Faced with this reality, health educators have developed survival strategies that read like a resistance handbook.
Make yourself robot-readable. The most successful educators have learned to speak fluent algorithm. They lead posts with phrases like “Sexual & reproductive health education” in big, clear text. They add “Educational content for 18+” watermarks to images. They include citations from medical journals in their captions. It shouldn’t be necessary, but it helps both AI and human reviewers understand context.
Dr. Jennifer Martinez, who’s built a following of 100,000 on TikTok despite multiple takedowns, shares her formula: “I literally start every video saying ‘I’m a board-certified gynecologist providing medical education.’ I wear my white coat. I have my diplomas visible in the background. I’m basically screaming ‘I’M A DOCTOR’ at the algorithm.”
Know your triggers. Through painful trial and error, educators have mapped the danger zones. On Meta and TikTok, any display of nipples or areolas—even in the most clinical context—is asking for trouble. Zoomed-in anatomical images get flagged more than full-body diagrams. Certain words in combination set off alarms: “teen” plus any anatomical term, “insertion” plus any body part, “pleasure” plus almost anything.
“I’ve started using euphemisms I swore I’d never use,” admits Chen. “I’m a medical doctor talking about ‘lady parts’ like I’m in middle school. But if that’s what keeps the content up, that’s what I’ll do.”
Separate your streams. Many educators now maintain different strategies for organic posts versus paid advertising. Ads face even stricter scrutiny, so successful campaigns use neutral creative that wouldn’t look out of place in a phone company ad, saving the actual health information for the landing page.
Document everything. The educators who successfully reverse takedowns are the ones who come prepared. They screenshot everything—the original post, the removal notice, the appeal rejection, timestamps, post IDs. They quote specific policy exceptions, like Meta’s stated allowance for “non-sexual nudity in educational contexts.” Some even cite Oversight Board decisions as precedent, essentially lawyering their way back online.
Build your reputation before you need it. Platforms are more likely to give established accounts the benefit of the doubt. Complete all verification steps. Maintain a history of uncontroversial content. Build your authority indicators. It’s unfair—new educators face higher barriers—but it’s reality.
The Human Cost of Digital Prudishness
Behind every removed post is someone who needed that information.
The teenager in a state with mandatory parental consent laws, trying to understand their options. The woman experiencing menopause symptoms, wondering if what she’s feeling is normal. The trans kid looking for accurate information about puberty blockers. The new parent worried about postpartum changes nobody warned them about.
When platforms remove legitimate health content, they’re not protecting these users—they’re abandoning them to misinformation, shame, and potentially dangerous alternatives.
“Young people don’t stop having questions just because we can’t answer them,” says Dr. Amanda Foster, who runs a reproductive health clinic in Texas. “They just go to worse sources. They find forums full of myths. They trust random influencers over medical professionals. They make decisions based on fear and misinformation instead of facts.”
The platforms know this. Their own research shows young users relying on social media for health information. TikTok regularly celebrates health education creators—when it’s not shadowbanning them. Meta publishes reports about combating health misinformation while simultaneously making it harder for health professionals to counter that misinformation.
The Path Forward (If There Is One)
This isn’t a story without hope, but the solutions require more than individual workarounds.
The Oversight Board has been surprisingly effective at pushing back, consistently ruling in favor of educational content and calling out Meta’s overreach. But it only covers Meta properties, moves slowly, and can’t address the daily grind of algorithmic censorship.
Some advocates push for regulatory intervention—laws requiring platforms to protect educational content or at least provide transparent, consistent policies. Others focus on platform accountability, using public pressure and media attention to force reversals of particularly egregious removals.
Technical solutions exist too. Better AI training could distinguish medical context from sexual content. Human reviewers with actual medical knowledge could handle health-related appeals. Verified health educators could receive different content moderation rules. None of this is impossible—it’s a matter of priorities and resources.
But until systemic change arrives, individual educators keep fighting their daily battles. They speak in code. They self-censor. They spread their content across platforms to avoid total deplatforming. They screenshot everything and prepare for the inevitable appeals.
Dr. Chen eventually got her account back after a journalist contacted Meta for comment. She’s rebuilt most of her following, though she’ll never know how many people needed her information during those dark weeks. She’s more careful now—uses more diagrams, fewer photographs, speaks in euphemisms she hates.
“I went to medical school to help people understand their bodies,” she says. “I didn’t expect to spend half my time fighting robots that think anatomy is pornography. But if that’s what it takes to get good information out there, that’s what I’ll do.”
She pauses, then adds something that captures the absurdity of our moment: “I just had to explain to a Silicon Valley algorithm that a cervix isn’t sexual. That’s not a sentence I expected to say in 2025.”
Your Survival Guide: What Actually Works
If you’re a health educator facing censorship:
- Front-load context: Start every post with “Medical education,” “Health information,” or similar. Make your intention unmistakable in the first three seconds of video or first line of text.
- Know the escalation ladder: In-app appeal → request human review → Oversight Board (if eligible) → media attention (if warranted). Don’t skip steps, but don’t give up either.
- Use the magic words: Quote specific policy exceptions. Meta allows “nudity in medical context” and “educational content about human anatomy.” TikTok permits “educational, documentary, scientific, or artistic content.” Use their own language against them.
- Build coalitions: Connect with other health educators. Share strategies. Amplify each other’s appeals. Platforms respond faster to coordinated pushback than individual complaints.
- Create backups: Cross-post to multiple platforms. Maintain an email list. Have a website. Don’t let any single platform control your ability to educate.
If you’re someone seeking health information:
- Follow multiple sources: Don’t rely on any single creator or platform. Good health educators will encourage this.
- Look for credentials: Real health professionals will display their qualifications prominently, especially given current censorship challenges.
- Support the educators you trust: Like, share, comment. Engagement helps combat shadowbanning. If someone’s content helps you, tell them.
- Speak up: When you see legitimate health content removed, report it. Tag the platform. The squeaky wheel gets the algorithm adjusted.
The Bottom Line
We’re living through a peculiar moment where the same platforms that revolutionized access to information are now gatekeeping basic health education. Where AI trained to protect children can’t distinguish between child exploitation and childhood development education. Where a medical degree means less than an algorithm’s split-second decision.
The educators adapting to this reality—speaking in euphemisms, fighting appeals, rebuilding after takedowns—are doing more than preserving their careers. They’re maintaining vital channels of health information in an increasingly censored digital world. They deserve better than fighting robots to teach reproductive health.
But until Silicon Valley figures out that cervixes aren’t inherently sexual, that breast exams aren’t pornographic, and that young people deserve access to accurate health information, the resistance continues. Post by carefully worded post. Appeal by exhausting appeal. One algorithmic battle at a time.
Because at the end of the day, this isn’t really about content moderation policies or AI capabilities. It’s about whose health matters, what information gets to spread, and whether we’ll let prudish algorithms determine what people are allowed to learn about their own bodies.
The answer from health educators is clear: not without a fight.
