AI Tool Diagnoses 18 Children With Rare Diseases Doctors Missed

Artificial intelligence has successfully diagnosed 18 children whose rare diseases had eluded traditional medical evaluation, according to new research from Boston Children's Hospital's center for rare diseases conducted in partnership with OpenAI.

The study demonstrates that commercially available AI tools can analyze patients' genomic data to pinpoint which genetic errors might be causing disease symptoms in pediatric cases. The findings, first reported by NBC News, represent a significant development in applying AI to rare disease diagnosis, where physicians often face challenges identifying the precise genetic mutations responsible for a child's condition.

How the AI analysis worked

The research team used off-the-shelf AI technology to examine genomic sequences from children whose conditions had stumped medical professionals. Rather than requiring specialized medical AI systems, the study showed that existing AI tools could be adapted to identify disease-causing genetic variants that human analysis had missed.

Rare diseases collectively affect millions of patients, but individual conditions may have only a handful of documented cases worldwide. This scarcity makes diagnosis particularly challenging, as physicians may never encounter certain genetic patterns during their careers. The genomic data for each patient can contain thousands of variants, requiring extensive analysis to determine which mutations are pathogenic versus benign.

Why it matters

For families navigating rare disease diagnoses, the typical journey involves years of uncertainty, multiple specialist consultations, and inconclusive test results. AI-assisted genomic analysis could dramatically compress this timeline, allowing children to receive targeted treatments sooner and families to access appropriate support services. The use of commercially available AI tools—rather than custom-built medical systems—also suggests this approach could scale more rapidly across healthcare institutions without requiring extensive specialized infrastructure.

Clinical implications

The Boston Children's Hospital research adds to growing evidence that AI can augment physician capabilities in complex diagnostic scenarios. Genomic medicine generates vast amounts of data that challenge human analytical capacity, making it a natural application area for machine learning systems.

However, the study focused on identifying genetic variants rather than replacing clinical judgment. Physicians still play essential roles in ordering appropriate tests, interpreting AI findings within the broader clinical context, and developing treatment plans based on confirmed diagnoses.

The collaboration between a major pediatric hospital and an AI company reflects the increasing intersection of healthcare institutions and technology firms in advancing medical diagnostics. As genomic sequencing becomes more routine in clinical practice, tools that can efficiently analyze this data will become increasingly valuable.

The research findings were first reported by NBC News correspondent Jared Perlo in June 2026.