Montefiore nurses fight datavant AI layoffs

Montefiore Medical Center's decision to lay off 12 utilization review nurses and replace their roles with Datavant AI technology has ignited fierce opposition from the New York State Nurses Association (NYSNA), which argues this move violates protections negotiated in the recent contract following a historic strike. NYSNA President Nancy Hagans emphasized that "Artificial intelligence should never replace real human caring from a nurse," underscoring the union's demand that AI serve as a support tool rather than a substitute for licensed nurses, with affected staff like Marilyn Shuler calling for direct nurse involvement in AI implementation to safeguard patient care and job security.

The conflict at Montefiore exemplifies broader labor concerns about AI-driven job displacement, surveillance, and contract violations in healthcare, with union leaders highlighting that rapid AI adoption threatens traditional nursing roles by shifting nurses into supervisory positions over machines rather than hands-on care. Rebecca Love noted the absence of federal or state safeguards, stating, "There’s no federal safeguards. There’s no state safeguards... And there’s no individual, by themselves, that is going to be able to withstand the pressure," prompting unions to push for AI-specific language in contracts to establish guardrails and ensure transparency in AI governance.

Unionized nurses are increasingly leveraging collective bargaining to secure meaningful involvement in AI policy decisions, recognizing that without such protections, employers can unilaterally impose AI-driven layoffs and surveillance measures. Contracts ratified recently at institutions including Montefiore now explicitly prohibit AI from replacing nurses or driving disciplinary and staffing decisions, reflecting a growing labor movement that demands transparency, safeguards against legal liability, and recognition of nursing judgment as essential to patient safety amid widespread distrust—60% of nurses surveyed by National Nurses United do not trust their employers to prioritize safety when deploying AI.

Nurses and clinicians at Montefiore and beyond express deep concerns that AI-driven clinical decision tools often overlook critical nuances essential for accurate patient care and insurance reimbursement, as highlighted by laid-off nurse Ajita Mathew who warned that AI automation misses subtleties in charts and claims. This disconnect is echoed in a National Nurses United survey where 69% of nurses reported that AI acuity scores frequently conflict with their professional judgment, raising alarms about the potential for AI to undermine nursing assessment and patient safety.

The rapid integration of AI without sufficient nurse training or involvement exacerbates liability risks and threatens the quality of care, with Charlotte Tschider cautioning that many nurses lack the preparation to properly oversee AI tools, potentially exposing both healthcare systems and individual practitioners to legal jeopardy. Compounding this, 64% of nurses feel unprotected legally if AI causes harm, and 35% report feeling pressured by employers to use these tools despite reservations, underscoring a troubling gap between AI deployment and frontline clinical readiness.

AI surveillance technologies, such as those used at Kaiser Permanente to grade empathy and call tone, have drawn sharp criticism from nurses who argue these tools erode compassionate care and professional judgment. Nurses circulated a petition stating “Trust nurses, not AI,” describing how AI-driven performance metrics—like strict call length limits—force them to truncate emotionally vital conversations, sometimes with dire consequences, such as a nurse staying over an hour on a call with a suicidal patient yet facing criticism for exceeding time limits.

Beyond immediate clinical impacts, experts warn that AI’s encroachment risks diminishing critical thinking and clinical skills over time, particularly among medical trainees who may become overly reliant on flawed AI tools. Simar Bajaj and Joseph Sakran caution that AI encourages 'cognitive offloading,' impairing brain activity during decision-making and potentially leading to a generation of clinicians who never fully develop essential reasoning and empathy. They advocate for structured limits on AI use in training, akin to pilots periodically flying manually, to preserve cognitive sovereignty and maintain human judgment at the heart of patient care.

Effective AI integration in healthcare hinges on robust governance frameworks that ensure data quality, seamless workflow integration, and continuous evaluation. Barry Mulholland of MBI Health emphasizes that “poor data doesn’t become reliable simply because it is processed by AI,” underscoring the primacy of data integrity over technology hype. Meanwhile, Steve Wightman from Access Health highlights the necessity of embedding AI tools directly into clinical records rather than relying on standalone applications, warning that without adequate funding and trust-level empowerment, these tools risk becoming burdensome add-ons rather than facilitators of care.

A growing disconnect exists between frontline clinicians’ rapid adoption of AI and the slow, cautious pace of institutional governance, creating significant risks around privacy, accountability, and patient safety. Nearly 86% of clinicians reportedly use AI frequently without employer guidance, often resorting to personal accounts that evade organizational oversight, as noted in recent European analyses. This 'shadow AI' phenomenon, described by Carebility CEO Jason Jones, exposes healthcare organizations to the 'mosaic effect'—where anonymized data can still be re-identified—highlighting urgent needs for comprehensive AI policies, cross-functional oversight, and rigorous vendor vetting to uphold HIPAA compliance and safeguard patient information.

Insufficient, non-specialty-specific training remains a critical barrier to realizing AI’s promise in healthcare, often leading to clinician frustration and underutilization of tools. KLAS data reveals fewer than 25% of clinicians feel adequately trained, with satisfaction declining when more than four AI tools are deployed without sustained, iterative education. Clinicians are not resistant to AI itself but to being handed unsupported technology, a gap exacerbated by budget cuts in digital training amid workforce shortages. This training deficit risks multiplying documentation burdens rather than alleviating them, underscoring the need for tailored, ongoing educational investments to ensure AI tools genuinely reduce workload and enhance patient care.

The rapid, often accelerated deployment of AI-driven clinical decision and documentation tools raises profound safety and ethical concerns, particularly when governance structures lag behind adoption. Kaiser Permanente’s use of AI surveillance to monitor nurse call times, for instance, pressures staff in ways that may compromise compassionate care, while union agreements limit nurses’ power to reject or suspend such systems. Moreover, AI-generated medical notes risk inaccuracies, fabricated information, and bias, with minimal federal oversight turning hospitals into unwitting laboratories for patient safety and billing integrity experiments. Experts like Thomas Tzimas stress the indispensability of human oversight, mandating clinician review and accountability to mitigate hallucinations and errors, while calls grow for transparent, evidence-based frameworks that include validation, bias testing, and clear liability standards.