QueerAF’s investigation finds that popular AI models frequently frame trans existence as a legal or medical issue, shaping how queerness is presented in outputs.

The research shows Trans+ topics are described as a mix of biological, psychological, and social factors, with responses often presenting competing viewpoints in a ‘both-sides’ framing.

Scholars and advocates call for inclusive training data, retraining of language models, and safeguards to protect marginalized communities, while acknowledging this work is complex and time-consuming.

The piece advocates for more trans-informed journalism and urges QueerAF membership to support reporting that centers lived experience and accuracy.

It emphasizes a commitment to human-led journalism, skepticism toward AI-sourced content, and a call for readers to join QueerAF to back responsible reporting.

Reliance on AI across apps risks entrenching censorship and erasing queer histories, potentially flattening nuanced experiences in archives and research.

Week’s issues include signs that AI dilutes Trans+ experiences, NHS Wales anti-trans complaint documents affecting surgery decisions, and concerns raised by an intersex open letter about medical information given to patients.

The framing is seen as a result of anti-trans policy language and media narratives that shape public understanding through AI-generated content.

Examples like Anthropic's Claude illustrate how even where risk is statistically small, the topic becomes a contested legal/political debate, revealing bias in framing.

There is concern that AI summarizations could dominate information channels and influence perceptions of trans rights and identities.

Ongoing debate about avoiding AI or pursuing bias-aware, inclusive training, with calls for regulatory safeguards and ethical AI practices.

The investigation notes frequent citations of legal cases and policy instruments (For Women Scotland, EHRC, Forstater v CDG) and the Cass Review, sometimes without critical methodological context.