Loading strategy tools...
The lawsuits over AI in health care have arrived — and California is the epicenter. From ambient scribes recording exam rooms to algorithms denying claims to pixels leaking patient data, this tracker maps every case, verdict, and binding rule, with a primary source on each. Built for the FQHC running these tools tomorrow morning.
18 cases tracked · 11 in California · Updated: Jun 3, 2026
18
Cases
11
In California
11
Active
7
CA laws
Why it matters: HIPAA may not reach the pixels on your public pages — but California's CIPA and CMIA absolutely do, at $5,000 per violation.
No California FQHC has been sued by name yet. But the ambient AI scribes (Abridge) that community health centers already run are named in two active California cases. This is coming — start with the risk self-check below.
The story
How we went from the first algorithmic-denial suits to the first CIPA jury verdict to AI-scribe lawsuits in the exam room. Tap any milestone for the detail and the source.
Does this apply to you?
Check the AI tools your clinic actually uses. We'll show which laws bind you, which live cases map to your exposure, and where you hold leverage.
Check the AI tools your FQHC uses. Nothing is saved — it all computes in your browser.
What does your clinic run?
Select your tools
Toggle the AI tools your clinic actually uses to see your exposure.
The math
CIPA sets statutory damages for every violation, for every patient. Move the slider to see why one chat widget or AI scribe can become existential exposure — then compare it to what California providers actually paid.
Why one chat widget or AI scribe is an existential risk: $5,000 per violation, per patient.
Theoretical statutory exposure
$250M
50,000 × $5,000 per violation (CIPA §637.2)
Suits almost never pay the full statutory max. Settlements typically land at a small fraction — perhaps $2.5M–$13M in this scenario. But even a fraction is existential for an FQHC, and the Flo Health verdict (Aug 2025) proved a California jury will impose CIPA liability.
What they actually paid
CA hospital, pixel settlement
CA community hospital — the FQHC-scale analog
National comparator (non-CA)
13.4M people — largest tracking settlement
The docket
Filter by category or by California. Every card links to a primary source and flags anything not yet verified against a court docket.
Washington v. Sutter Health; companion v. MemorialCare
The system: Abridge ambient AI scribe — records and transcribes clinical encounters, drafts notes, and writes them into the EHR.
Saucedo v. Sharp HealthCare
The system: Abridge ambient AI scribe — deployed ~April 2025 to record clinician–patient encounters and draft notes; audio allegedly sent to the vendor's cloud.
Estate of Gene B. Lokken v. UnitedHealth Group
The system: nH Predict — a NaviHealth (Optum) algorithm that predicts post-acute length-of-stay for Medicare Advantage members. Plaintiffs allege it cut off coverage in lieu of physician review, with a ~90% appeal-reversal rate.
Kisting-Leung v. Cigna Corporation
The system: PxDx ('procedure-to-diagnosis') — an automated system that flags claims against a list of 'acceptable' procedures and batch-denies them. Reporting found physicians denied 300,000+ claims over two months at ~1.2 seconds each.
Kaiser Permanente Tracking Technology Settlement
The system: Pixels and session-replay (Google, Bing, X, Adobe, Quantum Metric) on authenticated patient portals and apps — secure pages where members log in to view records or message doctors.
Up to $47.5 million
In re Meta Pixel Healthcare Litigation
The system: Meta Pixel — JavaScript tracking code on hospital/provider websites and patient portals that allegedly transmitted patient health information to Meta for ad targeting.
Frasco v. Flo Health, Inc. (Meta, Google, Flurry)
The system: A software development kit (SDK) embedded in the Flo period/fertility app that transmitted reproductive-health data to third parties.
Statutory exposure in the billions ($5,000/violation)
Popa v. Microsoft Corp.
The system: Website analytics / session-tracking (analyzed for Article III standing).
In re Otter.ai Privacy Litigation (Brewer / Walker / Theus)
The system: Otter.ai ambient transcription/notetaker — auto-joins meetings, records audio, creates voiceprints, and trains models on recordings.
Eisenhower Medical Center Pixel Settlement
The system: Meta Pixel + Google Analytics embedded without consent, disclosing patient info to third parties.
$875,000
MarinHealth Medical Center Pixel Settlement
The system: Meta Pixel on a regional Northern California community hospital's website.
$3 million
Barrows v. Humana, Inc.
The system: nH Predict — the same NaviHealth post-acute algorithm at issue in Lokken; Humana is also a NaviHealth client.
Couture v. OpenAI
The system: Meta Pixel + Google Analytics embedded in the ChatGPT.com interface.
CVS Health / Aetna post-acute denial investigations
The system: CVS/Aetna post-acute denial tooling. A 2024 U.S. Senate report documented a CVS 'Post-Acute Analytics' initiative projected to save $77.3M by using analytics to cut skilled-nursing spend.
Doe v. Eating Recovery Center, LLC
The system: Meta Pixel on an unauthenticated treatment-information website.
Garcia v. Character Technologies, Inc. (and Google)
The system: Character.AI conversational chatbot/LLM (not a medical device).
Dinerstein v. Google, LLC & University of Chicago Medical Center
The system: Training data — the medical center gave Google years of de-identified EHR records to train predictive-care algorithms.
In re Advocate Aurora Health Pixel Litigation
The system: Meta Pixel + Google trackers, including on the MyChart portal.
$12.225 million
Epic Sepsis Model external-validation controversy
The system: Epic's proprietary EHR-embedded sepsis early-warning score, deployed at hundreds of US hospitals (the dominant FQHC EHR via OCHIN Epic).
Each case is graded by category (AI scribes, coverage denial, privacy/pixels, clinical AI), jurisdiction, and severity. 'Newly filed' or 'In discovery' means the case is active. Every claim links to a primary source — a court docket, a regulator's release, or authoritative legal coverage — and anything we couldn't verify against a primary source is flagged. This is intelligence tracking, not legal advice.