detrans.ai: A Counter Narrative | Peter James Steven
26 February 2026
With Peter James Steven
Global
Peter James Steven built detrans.ai as a structured counter to the affirmation defaults embedded in mainstream AI models, drawing directly on detransitioner testimony and data. The episode examines what the chatbot's database reveals about transition and detransition patterns — evidence that European health regulators, from Sweden's medicines authority to the authors of the Cass Review, have identified as a critical and long-neglected gap in the clinical literature.
Peter James Steven is the creator of detrans.ai, an AI-powered chatbot built to surface the experiences and testimonies of people who have detransitioned — those who underwent medical or social gender transition and subsequently reversed course. In this episode of Beyond Gender, he explains the motivations behind the project, its technical architecture, and what the data gathered so far reveals about patterns in transition and detransition. The chatbot was designed explicitly as a counter-narrative to the outputs generated by mainstream AI systems such as ChatGPT, which Steven argues reflect and reinforce gender-affirmation ideology. Where dominant AI models tend to validate and encourage gender transition, detrans.ai draws on a curated body of detransitioner testimony and research to present a more cautious and evidence-grounded picture. The distinction matters: AI systems increasingly shape how vulnerable people — including minors — first encounter information about gender identity and medical options. The episode gives considerable attention to the statistics emerging from the project's database. Patterns in why people transition, what factors correlate with detransition, and how detransitioners describe their experiences in retrospect offer a form of real-world evidence that has been largely absent from clinical literature — or actively contested. Detransitioners have historically reported difficulty being heard within healthcare systems designed around an affirmation-only model. Steven's project represents one attempt to aggregate and systematise that experience at scale. The discussion also touches on the role of pronouns in shaping gender identity, the personal stories surfacing through the platform, and broader demographic trends among those who transition and later detransition. These are not peripheral anecdotes. Regulatory bodies across Europe — from Sweden's Medical Products Agency, which restricted puberty blockers and cross-sex hormones for minors, to the independent Cass Review, which prompted sweeping changes to NHS gender services — have pointed explicitly to the inadequacy of the evidence base underpinning paediatric gender medicine. Detransitioner outcome data is one of the documented gaps that reviewers have identified as requiring systematic collection. For researchers, clinicians, and policymakers working in this space, a searchable, structured dataset of detransitioner experience carries real evidentiary weight. European health authorities have increasingly acknowledged that long-term outcome data — including data on those who come to regret transition — has been missing from the evidence base used to justify clinical protocols over the past two decades. Efforts to compile and systematise that data, whether through formal registries or tools like detrans.ai, address a need that official reviews have now put on record. Steven also reflects on his personal motivations for building the tool and on the importance of open dialogue in a field where dissenting voices have frequently been marginalised. The conversation illustrates how civil society and technology can contribute to evidence generation in spaces where formal research has been slow, contested, or shaped by ideological pressures rather than clinical outcomes.
