Algorithms in the Closet: Why Are LGBTQ+ Creators Once Again Paying the Price for Big Tech's Moderation Failures?

 

I live in a part of the world where creating content is king no matter what your sexuality may be.

It is becoming an increasingly familiar story for LGBTQ+ creators. Years of work disappear overnight. Entire YouTube channels vanish without warning. Videos are removed. Monetization is stripped away. Appeals are rejected or met with silence. Then comes the same explanation many creators say they receive: an automated system has determined their content violates policies relating to nudity or sexual content.

 

For those whose livelihoods depend on digital platforms, this is not merely an inconvenience. It can mean the loss of years of creative work, thousands of subscribers, advertising revenue, sponsorship opportunities and, in some cases, a significant portion of their income. A channel is not simply a social media account. For many, it is their business.

 

To be clear, there is no credible evidence yet that YouTube is deliberately conducting a coordinated campaign to remove LGBTQ+ creators simply because they are LGBTQ+.

 

However, that does not diminish the seriousness of what many creators say they are recently experiencing. Reports of wrongful strikes, demonetization and channel removals have become increasingly common, and many affected creators believe automated moderation systems are incorrectly flagging perfectly legitimate LGBTQ+ content as sexual or otherwise inappropriate.

 

That distinction matters. Intent may not be discriminatory, but outcomes can still disproportionately harm a particular community.

 

The timing is difficult to ignore. YouTube has significantly expanded automated enforcement across the platform, relying more heavily on artificial intelligence to detect potential violations before human review. While creators across many genres have reported erroneous strikes, LGBTQ+ creators have become particularly vocal about what they see as algorithms that fail to distinguish between sexual content and LGBTQ+ identity or expression.

 

Travel creators documenting Pride festivals. Couples sharing their daily lives. Gay men enjoying a beach holiday. Educational discussions about LGBTQ+ issues. Lifestyle content. None of these are inherently sexual. Yet many creators claim their content is repeatedly caught in moderation systems that appear unable to make that distinction.

This is not an entirely new problem.

 

Back in 2017, YouTube acknowledged that a technical issue with its Restricted Mode incorrectly filtered hundreds of thousands of LGBTQ+-related videos, effectively hiding them from many users. The company eventually corrected the issue, but it demonstrated something many LGBTQ+ creators had already suspected: algorithms can reinforce bias even when discrimination is not intentional.

 

Nearly a decade later, many creators fear history is repeating itself.

 

The concerns extend beyond YouTube.

 

Across Instagram, Facebook and other Meta-owned platforms, LGBTQ+ creators have spent years battling opaque moderation systems that frequently confuse identity with explicit content. Images of men at beaches, shirtless photographs taken in perfectly ordinary settings, or promotional material for LGBTQ+ venues can trigger removals or account suspensions despite containing nothing that would reasonably be considered pornographic.

 

My own experience reflects that frustration.

 

Over more than a decade managing LGBTQ+ social media accounts, including Gay French Riviera, I have seen my accounts suspended multiple times and anywhere from 3 to 30 days with appeals going nowhere. But at least Meta generally provides some explanation, even if that explanation often appears disconnected from reality.

 

As an example, a photograph of a man standing on a beach on the French Riviera somehow becomes problematic despite beaches being full of people wearing swimsuits every single day.

AI-generated promotional images featuring shirtless men—even when they are not touching or posed suggestively—have also triggered restrictions. Some LGBTQ+ venues have become almost impossible to promote because automated systems appear to treat their imagery differently than comparable mainstream tourism content.

 

Even language itself seems to create problems. 

 

Many creators now warn each other against combining perfectly ordinary words such as “travel,” “beach,” “gay” and “men” in titles, descriptions or hashtags out of fear that algorithms may incorrectly classify their content.

 

Whether every one of these claims can be independently verified is beside the point. The fact that creators are modifying perfectly legitimate speech to avoid invisible moderation systems demonstrates how deeply trust has eroded.

Adding insult to injury is the lack of transparency.

 

When a creator spends years building an audience, carefully following platform guidelines, investing countless hours filming, editing and uploading videos, they deserve more than an automated email generated by artificial intelligence. They deserve to know precisely what policy was violated, exactly which content triggered enforcement and the opportunity for meaningful human review.

 

Instead, many describe appeal systems that feel equally automated, leaving creators trapped in a bureaucratic loop where no one appears willing—or able—to explain what actually happened.

 

Organizations such as GLAAD have also criticized YouTube’s moderation practices, arguing that recent policy changes have contributed to a less safe online environment for LGBTQ+ users and creators. While GLAAD has not alleged a deliberate purge of LGBTQ+ accounts, its concerns echo what many creators have been saying for months: automated moderation systems are producing outcomes that disproportionately affect LGBTQ+ communities.

 

Artificial intelligence undoubtedly plays an important role in moderating billions of pieces of content uploaded every day. No reasonable person expects every video to receive immediate human review. But when AI systems repeatedly struggle to distinguish between LGBTQ+ identity and sexual content, the technology is no longer simply imperfect—it begins producing systemic consequences.

 

These are not isolated technical glitches. They become barriers to participation, visibility and economic opportunity.

 

For LGBTQ+ creators, visibility has never been guaranteed. It has been fought for over decades. Social media platforms often present themselves as champions of diversity during Pride Month, wrapping logos in rainbow colours and celebrating inclusion through carefully crafted marketing campaigns.

 

Those commitments ring hollow if the very systems powering these platforms continue to suppress, demonetize or remove legitimate LGBTQ+ voices through flawed automated enforcement.

 

If artificial intelligence is making these mistakes, then the companies deploying that technology have a responsibility to fix them.

 

That responsibility includes meaningful transparency, faster access to human review, detailed explanations when enforcement occurs, independent auditing of AI moderation systems for potential bias, and public reporting on whether certain communities are disproportionately affected by automated decisions.

 

Until that happens, every unexplained suspension sends the same troubling message to LGBTQ+ creators: your years of work, your audience and your livelihood can disappear at the click of an algorithm—with little explanation and even less accountability.

 

Some affected creators have suggested filing complaints through the Digital Services Act framework in the European Union, including with their national Digital Services Coordinator, as one avenue for challenging moderation decisions where applicable. Whether that process proves effective remains to be seen.

 

What is already clear is this: if Big Tech wants creators to trust artificial intelligence with decisions that affect careers and communities, then AI cannot become an unaccountable black box. Technology should protect free expression, not repeatedly place marginalized communities in the position of having to prove that simply existing, travelling, speaking or sharing their lives is not a policy violation.

 

This seems to be just one more battle for us to fight. But fortunately for us, this ain’t our first time at the rodeo.