Montefiore lays off 12 nurses, replaces them with Datavant AI

A dozen utilization review nurses at Montefiore hospital in the Bronx were told they were being replaced by AI-powered software, and their union is calling it a contract violation. According to The Guardian, the 12 nurses received 45-day layoff notices on May 28, months after a January 2026 strike across New York City hospitals produced new contracts that included safeguards against AI displacing clinical staff.

Utilization review is the paperwork side of care, reading patient charts and communicating with insurance companies over coverage, and it is exactly the kind of work hospital systems are eager to automate. The vendor here, according to a NYSNA press release, is Datavant, a private-equity-backed technology company that also has partnerships with Palantir. Marilyn Shuler, RN, who spent 39 years at Montefiore, was among those laid off. Her ask, as reported, was narrow rather than absolutist: keep a licensed nurse on the final review and use the software to support the humans rather than replace them.

Montefiore disputes the framing. Joe Solmonese, the hospital's senior vice-president for government relations and strategic communications, called the union's claims 'inaccurate and misleading' and described the affected work as a 'nonclinical program involving paperwork.' That is the honest tension in the story: the nurses argue that even the paperwork step is clinical judgment, because a licensed nurse's read of a chart is what shields a patient from a wrongful insurance denial, and the hospital's position is that this is administrative work that automation can handle.

The caveat worth being honest about is that the reporting is single-sided in places. You have the union's version of the contract language and Datavant's role, and a short statement from the hospital rejecting the characterization. What the reporting does not give you is the specific Datavant product being deployed, how its outputs are reviewed, or the audit trail. That matters, because the difference between a tool that flags cases for a nurse and a system that issues the final review is the whole argument.

For hospital operators watching this, the useful signal is not whether AI can do utilization review. It clearly can do some of it. The signal is that AI safeguards written into a fresh union contract are already being tested inside the first year, and the next round of bargaining across the sector is going to be much more specific about what 'replacement' actually means.