GLAAD President and CEO Sarah Kate Ellis said Wednesday that AI bias puts LGBTQ+ people at risk [1].
This warning highlights the potential for artificial intelligence to scale systemic prejudice, turning flawed training data into automated misinformation that can impact real-world safety and civil rights.
Speaking at the Axios AI+ NY summit in New York City on June 3, 2026 [1], Ellis said that AI systems trained on biased data can reinforce harmful stereotypes. She said that these technical flaws do not exist in a vacuum but can actively spread misinformation about the LGBTQ+ community [2].
Ellis said the urgency of addressing these algorithmic failures is necessary to prevent further marginalization. The CEO said that the current landscape of digital attacks is evident to those paying attention.
"I think you'd have to be asleep at the wheel to not see the attacks that are happening against ..." Ellis said [1].
The concerns raised at the summit center on how large language models and generative AI ingest vast amounts of internet data. Because this data often contains historical biases and hate speech, the resulting AI outputs may mirror those prejudices, creating a feedback loop of discrimination [2].
Ellis said there must be greater accountability in how these models are developed and monitored. Without intervention, the organization said that AI could become a tool for automating the erasure or vilification of LGBTQ+ identities [1].
“AI bias puts LGBTQ+ people at risk”
The intersection of generative AI and social bias suggests that technical efficiency does not equal neutrality. When AI models are trained on uncurated web data, they risk codifying existing societal prejudices into software. For marginalized groups, this means the risk of 'algorithmic discrimination,' where AI-driven tools for hiring, moderation, or information retrieval may systematically disadvantage or misrepresent them based on biased training sets.
