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

Artificial intelligence is amplifying anti-LGBTQ bias, misinformation, and discrimination in ways that can affect everything from healthcare and employment to housing and privacy, according to a new report from advocacy organization GLAAD.

Released on Wednesday, the report titled “Build for Everyone: A Framework for LGBTQ Representation and Safety in AI,” argues that LGBTQ safety should be treated as a core requirement of responsible AI development.

GLAAD warns that AI trained on biased or incomplete data can reinforce stereotypes, suppress LGBTQ voices, expose users to privacy risks, and produce discriminatory outcomes as the technology becomes increasingly embedded in everyday life.

“AI is a civil rights issue,” GLAAD President and CEO Sarah Kate Ellis wrote in the report. “Neutrality is no longer an option. To build AI that is ethical, inclusive, and responsible, tech leaders must proactively embrace intentional practices to create safe products.”

Areas of concern, the study said, include biased training data, anti-LGBTQ misinformation, discriminatory outcomes in predictive AI systems, content moderation failures, and privacy risks, arguing that AI systems trained on incomplete or inaccurate information about LGBTQ people can reinforce stereotypes.

Ellis argued that responsible AI is best for business and a requirement for future-proofing AI companies. “More than 20 percent of Gen Z is LGBTQ,” she wrote. “These are your future employees and consumers.”

According to a 2023 study by advisory and investment firm LGBT Capital, the global buying power of LGBTQ people is $4.7 trillion, with that number estimated to reach $33 trillion by 2030.

“To put that in perspective, if we were a country, we would be the 4th largest economy in the world,” Ellis wrote.

The report comes amid an ongoing debate over AI bias. In May, researchers found leading AI models consistently favored Catholicism while responding less favorably to Jehovah’s Witnesses, atheism, and agnosticism.

Earlier this month, former xAI engineer Devin Kim sued xAI and SpaceX, alleging he was fired after warning that Grok lacked adequate safeguards against misinformation and bias. Meanwhile, the Elon Musk-led xAI is fighting a legal battle against Colorado over a state law requiring companies to assess and reduce discrimination risks in AI systems used for decisions involving housing, employment, and lending.

GLAAD argues the consequences extend beyond chatbot conversations and image generators.

“While not specific to LGBTQ people, we must also mention other emerging challenges as AI development and adoption progresses,” the study said. “These can include model hallucinations or sycophantic behavior that generate misinformation, including about consequential topics such as health or elections.”

Those concerns become more significant, the study said, as companies push AI agents capable of performing tasks with limited human oversight. GLAAD warned that autonomous agents could inherit existing biases and automate discriminatory outcomes, such as excluding LGBTQ-affirming healthcare providers from search results or making incorrect assumptions about users’ identities.

To prevent those risks from becoming further embedded in AI systems, GLAAD called on developers to improve LGBTQ representation in training data, strengthen privacy protections, maintain human oversight of moderation systems, and work more closely with advocacy groups. The report also calls for stronger industry accountability and regulatory oversight.

“Failure to account for LGBTQ experiences and issues in training data, product design, and governance can result not only in harm to marginalized communities but also in inaccurate, lower-quality products that may undermine user trust in a growing demographic,” the study said.