A federal court in Chicago is now the venue for one of the most consequential tests yet of how far an AI health company can go in monetizing genetic data. On April 15, 2026, the U.S. District Court for the Northern District of Illinois consolidated a wave of class action lawsuits against Tempus AI into a single complaint, Farrier et al. v. Tempus AI, Inc., accusing the company of disclosing patients’ genetic information to pharmaceutical and biotech partners without their knowledge or written consent.
How Tempus Built Its Genetic Data Business
Tempus AI markets itself as a precision-medicine company that applies artificial intelligence to genomic and clinical data to help doctors choose cancer treatments and to help drugmakers design trials. Its data trove grew substantially in February 2025, when it closed a roughly $600 million acquisition of Ambry Genetics, a clinical genetic testing lab that processes hereditary cancer and other genetic tests for patients across the country. That deal gave Tempus a much larger pool of raw genetic sequencing data to feed into its AI models and, according to the lawsuit, to package for sale.
The complaint alleges that Tempus disclosed genetic data through licensing and research agreements with a roster of major pharmaceutical and biotech companies, including AstraZeneca, Bristol Myers Squibb, Pfizer, GlaxoSmithKline, Incyte, Servier, Aspera Biomedicines, and Whitehawk Therapeutics, in deals plaintiffs say collectively total $1.1 billion.
The Legal Theory: Genetic Data Cannot Be Anonymized
The core of the plaintiffs’ argument is not simply that Tempus sold data, but that the data it sold was never truly anonymous to begin with. The complaint states that genetic information “cannot be deidentified because such data serves as an inherently unique biomarker,” a claim that echoes a growing body of research showing that genomic sequences function almost like a fingerprint that can, in combination with other datasets, be traced back to a specific individual. Plaintiffs argue this makes Tempus’s data-sharing practices a violation of the Illinois Genetic Information Privacy Act, along with related state statutes that require explicit written consent before genetic material is shared with third parties.
A Pattern Playing Out Across Healthcare AI
The Tempus case did not emerge in isolation. Legal trackers report that data breach and privacy-related federal filings hit record levels in the first half of 2026, with more than 4,200 privacy filings logged in the first two quarters alone, according to litigation analytics reviewed by legal publications. Parallel cases have targeted a range of AI-adjacent healthcare practices this year, including a lawsuit against Mayo Clinic alleging its AI systems prioritized speed over accuracy and put patient privacy at risk, and separate suits against Sutter Health and Memorial Healthcare Services over ambient AI tools that record patient-clinician conversations without adequate consent. Medtronic, meanwhile, faces its own proposed class action after hackers obtained more than nine million records containing sensitive personal and health information in an April 2026 breach.
Industry Defenders See Innovation at Stake
Tempus and its allies in the health-data industry argue that large, well-curated genomic datasets are exactly what is needed to train AI models capable of matching patients to the right cancer therapy, and that overly restrictive consent requirements could choke off research that has already produced faster, more targeted treatments. From that vantage point, data-sharing agreements with drugmakers are not a betrayal of patients but the mechanism by which genomic AI turns into approved therapies, and litigation risk of this scale could push genomics companies toward more conservative, less useful de-identification practices that slow down legitimate research.
Privacy Advocates Say the Consent Was Never There
Patient advocates and privacy attorneys take the opposite view, arguing that consent buried in a testing lab’s terms of service is not the same as a patient knowingly agreeing to have their genetic code licensed into a nine-figure pharmaceutical deal. They point out that people who submitted a cheek swab to Ambry Genetics for a hereditary cancer screening, often at a moment of acute personal anxiety about a family history of disease, had no reasonable way to anticipate that their sample would end up feeding commercial AI partnerships worth over a billion dollars in aggregate. For this camp, the Tempus suit is less about blocking innovation and more about forcing companies to build consent into the business model from day one rather than after a lawsuit forces the issue.
What the Case Means for AI and Health Data Going Forward
Regardless of how Farrier v. Tempus AI resolves, the consolidation of these claims into a single federal complaint signals that courts are treating AI-era genetic data disclosures as a distinct and serious category of privacy harm, not just a variant of ordinary data breach litigation. Plaintiffs are seeking a jury trial, damages, and injunctive relief that would require written consent before any future data sharing. If they succeed, it could force Tempus and comparable genomics-AI firms to overhaul how they structure licensing deals with pharmaceutical partners, and it would give plaintiffs’ attorneys a template to challenge similar arrangements elsewhere in the healthcare AI industry. For patients, the case is a reminder that the fine print attached to a routine genetic test can end up defining the boundaries of a much bigger AI data economy than most people realize they are participating in.
Photo: sergeitokmakov / PIXABAY via Pixabay
