The American LGBTIQ organization GLAAD is sounding the alarm: Artificial intelligence (AI) threatens to reproduce and amplify the same biases, misinformation, and hate campaigns that have long plagued traditional social media. This warning comes from GLAAD’s report “Build for Everyone: A Framework for LGBTQ Representation and Safety in AI.” The document makes a stern case that unregulated or poorly trained AI models pose real dangers to the queer community—from algorithmic discrimination in credit decisions to dangerous medical misinformation.
For example, in Meta’s new Llama-4 model, “conversion therapies” for homosexuality were reportedly recommended—the guidance appearing when answering a question from the parent of a newly out child who was seeking information. Medical organizations agree that this kind of “treatment” is neither effective nor safe and that it can push individuals toward depression or even suicide.
Additionally, automated systems that lack understanding of context tend to over-block legitimate queer content and voices or hide them behind shadow bans. Predictive AI systems used in sensitive areas such as banking, housing, or job applications threaten to systematically disadvantage queer people if they are fed biased data.
Drastic Concerns in the Trans Community
The level of worry within the community is enormous according to current data. A 2025 survey by LGBT Tech found that 73 percent of queer adults fear AI-driven misinformation. Among trans people, the figure rises to a alarming 89 percent.
When AI systems are built on data that frame queer identities as “abnormal” or rights as “controversial,” the consequences for daily life can be drastic. GLAAD President Sarah Kate Ellis commented: “Neutrality is no longer an option. AI systems trained on data that wrongly position LGBTQ lives as a ‘fringe’ or treat our equality as ‘controversial,’ or that fail to curb sophisticated misinformation about LGBTQ people, threaten our health, safety, and civil rights.”
GLAAD: Accountability Over “Algorithmic Monoculture”
The core problem, according to GLAAD, is that a handful of American tech giants form the foundation for almost all downstream apps. This creates an “algorithmic monoculture.”
GLAAD urged developers to implement continuous human oversight and to keep their training data up to date in order to keep pace with the evolution of online hate. After all, AI also holds enormous potential for the LGBTQ community—such as providing a safer space to practice coming out or helping parents find reliable information. For that potential to be realized, companies must act now: “If AI fails LGBTQ people, it fails everyone. Companies must keep working to improve AI fairness, counteract bias, and strengthen the safety and quality of AI products for LGBTQ people and for everyone.”
