detrans.ai: A Counter Narrative | Peter James Steven
Peter James Steven built detrans.ai because he believed detransitioners deserved a presence in the AI conversation that mainstream tools have consistently denied them. In this episode, Stella, Mia, and Bret dig into the project's origins, its technical design, and the patterns that have emerged from its database — and explore what it means to build an AI tool that treats affirmation ideology as a question rather than a given.
Peter James Steven is the developer behind detrans.ai, a chatbot built specifically to surface the experiences of people who have detransitioned. Most mainstream AI tools default to affirmation-friendly frameworks, reflecting the institutional consensus that has dominated gender medicine for the past decade. Peter set out to build something different: a tool trained on the accounts of those who transitioned and later changed course, giving that population a presence in the AI conversation that they have rarely had anywhere else. Stella O'Malley, Mia Hughes, and Dr Bret Alderman bring exactly the right mix of expertise to this conversation. A practising psychotherapist, a gender medicine researcher, and a psychologist between them, the hosts can move fluently between the technical questions — what data shapes the model, how it was built, which design choices were made along the way — and the wider clinical and ethical implications of a project like this. The result is a conversation that is genuinely illuminating rather than simply promotional. Peter speaks openly about the personal connections that drew him to the project, and that dimension lifts the episode above a simple product demonstration. It is a conversation about why the silence around detransition has been so durable, and what it means for someone outside the established institutions to decide that filling that silence is worth their time and effort. A substantial part of the discussion focuses on what detrans.ai's database has actually revealed: patterns in who transitions, who later detransitions, and the reasons they give for both decisions. These statistics rarely appear in clinical guidelines or public policy, and the hosts are well placed to interrogate them — to ask what the numbers do and don't tell us, and where the gaps in the evidence remain. The conversation also covers ground that mainstream discussions tend to avoid: the relationship between pronoun adoption and the entrenchment of a gender identity, the stories detransitioners tell about their sexuality, and what those accounts suggest about the underlying drivers of transition decisions. Stella and Dr Alderman, drawing on their clinical backgrounds, are particularly sharp in pressing on these questions. The episode makes a larger argument: that AI reflects the assumptions baked into its training data, and that those assumptions are not neutral. Most AI models have absorbed affirmation ideology as a default position. Peter James Steven built something that treats that ideology as a variable rather than a given — and this conversation is the right place to hear why that distinction matters.
Watch and subscribe on YouTube
Open this episode on the Beyond Gender channel
