Most clinical AI today is activated, not procured.
EHR vendors have not published hallucination rate, error type, severity, or risk disclosure for the summary and scribe features your hospital is already running. Idem Labs gives health systems, malpractice counsel, and malpractice carriers the independent read on what those features are doing in the medical record.
Default-on, validation-off.
About 70-80% of US health-system AI summarization arrives as a default activation, not a vendor selection. Epic's chart summarizer alone is invoked over 16 million times per month across its customer base, nearly 3x growth since November 2025. By the end of 2025, roughly half of US non-federal acute care hospitals were using EHR-integrated generative AI in clinical care.
In May 2026, Ontario's Auditor General audited the 20 AI scribes approved for clinical use in the province. All 20 had inaccuracies. 9 fabricated content. 12 captured a different drug than the one prescribed. 17 missed important mental-health details.
That evidence opens the same question for three different audiences. The hospital asks what its governance committee is on the hook for. The malpractice attorney asks what the AI actually generated at the time of the encounter. The malpractice carrier asks what exposure is sitting in the book of business at renewal.
Who we work with
Post-deployment audits of EHR-native AI features in production. Audit-defensible dossiers addressed to AI governance committees, CMIOs, and General Counsel.
Evidence-grade reconstruction of AI behavior in the medical record for in-house General Counsel and outside trial counsel. Risk assessment, vendor contract review, regulatory readiness, expert designation, and deposition support.
Portfolio-level AI exposure assessments for underwriting and renewal. Matter-level evidence reviews for claims handling and reinsurance documentation.
Who we are
Idem Labs was founded in 2026 by Matt Allison and Josh Ward Ph.D.
Matt Allison directs AI Products and Regulatory Strategy at Quest Diagnostics, where he launched the company's first patient and provider facing GenAI products and built the evaluation, red team validation, and post-launch observability programs that govern them. He spent twenty years before that in clinical quality and audit, including HITRUST certification, NCQA Population Health accreditation, and audit oversight of Novo Nordisk's Wegovy and Ozempic patient support programs.
Josh Ward earned his PhD in Statistics from UCLA, where his research focused on the design and analysis of trustworthy AI systems. He has worked as an AI researcher at Amazon, the Food and Drug Administration, Accenture, and the University of Utah School of Medicine, designing AI systems that serve millions of users across consumer search, federal regulatory infrastructure, and clinical decision support.
Read the evidence.
Then let's talk.
The June 2026 white paper documents what EHR-native AI features are doing in the medical record and what your organization is on the hook for.