OpenAI: AI Helps Physicians Diagnose Rare Genetic Diseases in Children

OpenAI’s reasoning model — a system that solves problems step by step, as opposed to conventional models that generate answers in a single pass — identified 18 new diagnoses in previously unsolved cases of rare genetic diseases in children.

Collaboration with the Medical Community

The project was developed in collaboration with physicians and research institutions. Rare genetic diseases are particularly challenging to diagnose because they affect a small number of patients and physicians rarely encounter them in clinical practice. That very absence of an experiential base makes AI tools potentially valuable: they can search medical literature and symptom patterns far faster and more broadly than any individual specialist.

What 18 Diagnoses Mean

Each of the 18 identified diagnoses relates to a case that had previously gone unanswered, which for affected families can mean the end of years of diagnostic uncertainty. Rare genetic diseases by definition affect fewer than 1 in 2,000 people, and globally more than 7,000 such conditions are known — the majority of which have no approved therapy.

Limits and Caution

OpenAI emphasizes that AI in this context acts as an assistive tool for physicians, not a replacement for clinical judgment. The specific names of institutions, the model name, and broader metrics have not been publicly disclosed at this stage. Independent clinical validation of the findings will be crucial for assessing the true scope of this approach.