detrans.ai: A Counter Narrative | Peter James Steven
26 February 2026
With Peter James Steven
Global
Peter James Steven created detrans.ai, an AI chatbot trained on detransitioner testimonies, to offer a counter-narrative to the affirmation-only assumptions built into mainstream tools like ChatGPT. The episode examines what the project's growing database reveals about why people pursue transition and why some come to regret it — evidence that Irish clinical services, which do not systematically track detransition outcomes, have largely neglected to gather themselves.
Peter James Steven is the creator of detrans.ai, an AI chatbot built specifically to surface the experiences of people who have detransitioned after social or medical gender transition. Where mainstream AI tools such as ChatGPT are trained on data that reflects affirmation-first ideology, detrans.ai draws on a curated database of detransitioner testimonies and associated research, enabling it to respond to questions about gender identity, transition regret and psychological wellbeing in ways that dominant AI platforms typically will not. The episode traces the personal motivations behind the project — connections to people affected by detransition that prompted Steven to ask why their stories were being systematically excluded from the public conversation. He walks through the technical decisions involved in building the tool, including the choice of underlying AI model and the ethical considerations of handling sensitive personal testimony responsibly. A central focus is the statistical picture the detrans.ai database has assembled. The data reveal patterns in why people pursued transition in the first place and which factors are most commonly associated with later detransition. These findings carry particular weight in Ireland, where the Health Service Executive's National Gender Service operates under affirmation-aligned protocols and where systematic outcome tracking for patients who later regret transition is essentially absent. The Cass Review, published in 2024, identified serious weaknesses in the evidence base underpinning gender-affirming care across the English-speaking world — weaknesses that Irish clinical practice has not yet fully reckoned with. Steven and the hosts examine how pronoun use can entrench rather than simply reflect gender identity, and what detransitioner accounts reveal about the long-term psychological consequences of early social and medical transition. This is directly relevant to Ireland's policy environment: the Gender Recognition Act 2015 introduced legal gender self-declaration for adults with no medical involvement, and recurring proposals to extend comparable access to minors have not been accompanied by any parallel investment in long-term outcome research. One of the episode's more striking points is that detransition rates are not tracked by Irish clinics or the HSE as a matter of routine. People who come to regret transition may find little formal support from the system that oversaw their original pathway. Steven's project represents a form of evidence-gathering — imperfect but earnest — that official health institutions have declined to undertake. The argument implicit in detrans.ai is that voices excluded from institutional data do not disappear; they accumulate, and they deserve the same consideration as the voices used to justify transition in the first place. The conversation closes with reflections on sexuality and identity, noting that confusion about sexual orientation appears repeatedly in detransitioner accounts. For Ireland, where same-sex relationships are fully legally recognised and socially accepted, the possibility that some gender-dysphoric young people were same-sex attracted individuals who might have benefited from honest, exploratory psychotherapy rather than medical transition is a question that practitioners and policymakers cannot continue to set aside.
