Allina Physicians Strike Over AI Control in Patient Care and Billing

About 150 unionized physicians at Allina Health's Mercy and Unity hospitals in Minnesota staged a four-day strike last week in a labor dispute centered not on pay or staffing ratios but on who controls how artificial intelligence is used in diagnosis, care decisions, and billing—the most publicly visible labor action over clinical AI governance in U.S. healthcare history.

The strike, which ended after four days with negotiations resuming, reflects three years of failed contract talks between the Twin Cities Physicians union and Allina Health. At the center of the dispute: physicians say management is deploying AI tools in ways that affect care decisions without physician oversight, and they want contractual guarantees that doctors retain final authority over how those tools operate in their clinical workflows.

What the Physicians Are Demanding

The union's core AI-related demands, according to HIStalk, include:

• Physician approval before any new AI tool is deployed in clinical settings where it influences diagnosis, treatment recommendations, or billing
• The right to override AI-generated documentation without administrative penalty or productivity score impact
• Transparency on training data and accuracy metrics for any AI system that surfaces recommendations to physicians
• No AI-based performance monitoring that ties compensation to metrics derived from AI systems without physician input on those metrics

Allina Health's position, as reported, is that these demands would give the physician union effective veto power over technology adoption decisions that management views as operational, not clinical. The health system argues that AI tools currently deployed are assistive, not autonomous, and that existing oversight mechanisms are sufficient.

Why This Strike Is Different

Healthcare labor disputes involving technology are not new. Nurses have struck over staffing ratios driven by acuity scoring algorithms. Radiologists have clashed with administrators over AI diagnostic tools that affect read volumes and billing. But those conflicts have typically been about workload or compensation impact.

The Allina dispute is about governance—who decides whether a tool goes in at all, and what authority a physician retains once it is there. That is a harder question, and one that health systems have largely avoided negotiating.

The four-day duration was short enough that patient care disruption was managed through contingency staffing, but long enough to put the issue on the radar of national physician organizations. The American Medical Association (AMA) has been watching the outcome closely. So has the American College of Physicians.

The Underlying Tension in Clinical AI

The Allina dispute makes visible a tension that has been building across health systems since ambient AI scribes and AI-assisted diagnostic tools began scaling in 2023 and 2024.

On one side: hospital administrators and AI vendors frame these tools as reducing documentation burden, improving note completeness, and surfacing clinically relevant information faster. On the other: physicians who have seen AI-generated notes introduce errors that required correction, or who have been asked to sign off on AI outputs under time pressure that does not allow meaningful review.

The billing dimension adds another layer. As Blue Cross insurers reported this week, AI coding tools added $1 billion in excess charges across a two-year period by documenting secondary conditions that were not actively treated—a dynamic that puts physicians in the position of signing notes that may misrepresent the care they delivered.

When a physician signs an AI-generated note they did not fully review, the legal and professional liability is theirs, not the vendor's. Physicians are not wrong to want control over a process where they bear the downstream risk.

What Other Health Systems Are Watching

The outcome of the Allina negotiation will be studied by health system administrators and physician unions across the country for two reasons:

Precedent on AI governance clauses. If Allina agrees to contractual physician oversight of AI deployment, it will be the first U.S. health system to do so in a collective bargaining agreement. That makes it a template—or a warning, depending on which side of the table you sit on.

Defining "assistive" versus "autonomous." Neither the AMA's policy guidance nor existing federal regulations draw a clear line between AI that assists clinical decisions and AI that effectively makes them. A negotiated contract between Allina and its physicians would fill that gap in practice even without regulatory action.

What to Watch

The immediate question is whether Allina and the Twin Cities Physicians union reach a contract before a second strike authorization vote is held. The broader question is whether the AMA, which has issued guidance on AI in medicine but has not taken a position on collective bargaining over AI governance, weighs in before the next negotiation cycle. A ruling or formal guidance from the AMA or another national physician body would shift the balance of every future health system negotiation involving AI tools.