GLAAD Says AI Is Failing LGBTQ Users—And Warns the Risk Is Growing

The rapid advancement of artificial intelligence is reshaping how information circulates, but not always in ways that serve marginalized communities. GLAAD’s recent warning about AI’s failure to protect LGBTQ users underscores a growing concern: that rapidly evolving AI systems are not just reflecting societal biases but amplifying them, often with little oversight. This issue matters because AI now influences everything from healthcare recommendations to hiring decisions, meaning its failures don’t just stay online—they can directly harm lives. For LGBTQ individuals, already disproportionately targeted by online harassment and discrimination, the stakes are particularly high. The report suggests that AI’s tendency to prioritize engagement and efficiency over fairness can reinforce harmful stereotypes, misgender users, or even out them against their will—a privacy violation with potentially dangerous real-world consequences. This problem is rooted in how AI is trained. Most large language models are built using vast datasets scraped from the internet, which contain decades of biased, dehumanizing, or outright false content about LGBTQ people. Without careful curation, these biases seep into AI outputs, from chatbot responses that pathologize gender diversity to recommendation algorithms that suppress LGBTQ creators’ content. What’s less discussed is how this harm compounds for intersectional identities—users who are both LGBTQ and racial minorities, for example, may face compounded discrimination when AI systems conflate their identities with stereotypes. The big unanswered question is whether regulators, tech companies, or civil society can move fast enough to rein in these harms. GLAAD’s report calls for transparency in AI training data and stronger accountability measures, but enforcement remains uneven. Meanwhile, the AI industry’s push toward more personalized and immersive systems—like AI companions or deepfake technologies—risks creating new avenues for abuse. As these tools become more embedded in daily life, the gap between their promise and their peril will only widen unless proactive steps are taken. The challenge isn’t just technical but ethical: ensuring that AI serves all users, not just the most privileged ones.