Our Works

AI systems screening children for developmental conditions, authorizing treatments, or determining special education eligibility create liability that compounds across time. A false negative on autism screening at age 4 means missed early intervention. That error does not resolve at age 8. It propagates. When the parent learns their child was misrouted by an algorithm your model powered, the statute of limitations may not have started running. Pediatric classification errors have discovery timelines that outlast your product roadmap. Audience: Enterprise sales teams at foundational model companies selling into healthcare and EdTech verticals. Compliance officers at health systems deploying AI triage. Policy teams at companies whose models are being fine-tuned for pediatric use cases without their direct oversight. Why it matters now: Your acceptable use policy probably prohibits "medical diagnosis." It probably does not address administrative routing, eligibility determination, or triage prioritization. The distinction matters to your legal team. It will not matter to a plaintiff's attorney representing a family whose child was denied services by a system your model powered. Illinois HB 1806, the Therapy Resources Oversight Act, took effect August 2025. State-level pediatric AI regulation is arriving faster than your policy team is tracking it.