AI and Surveillance Are Undermining Nursing Care at Kaiser Permanente

The promise of artificial intelligence in healthcare has long been framed as a tool to reduce burden and improve outcomes. But for many nurses at Kaiser Permanente, the reality on the ground looks quite different. Instead of freeing them to spend more time with patients, new AI-driven systems and heightened workplace monitoring are adding layers of complexity, eroding trust, and in some cases, directly interfering with the quality of care they strive to deliver.

These concerns aren’t isolated complaints. They come from frontline staff who’ve seen algorithms flag routine variations in vital signs as emergencies, prompting unnecessary interventions that disrupt workflow and alarm patients. Others describe being constantly tracked through badge sensors and software that logs every mouse click, keystroke, and movement — not to support them, but to measure productivity in ways that feel punitive. One nurse described the sensation as “being watched not to help you do better, but to catch you slipping.”

At the heart of the issue is a growing disconnect between the intentions behind these technologies and how they’re implemented. AI tools designed to predict patient deterioration or optimize staffing often rely on incomplete or biased data. When the system gets it wrong — which nurses say happens frequently — it’s not the algorithm that faces consequences. It’s the nurse who must explain why they didn’t act on a false alert, or why they ignored a prompt that turned out to be noise.

Surveillance, meanwhile, has crept into nearly every corner of the workday. Break times are monitored. Bathroom trips are logged. Even the time spent charting — already a significant burden — is scrutinized for efficiency. Nurses report feeling pressured to rush through documentation or skip steps to meet invisible benchmarks, all while knowing that shortcuts could compromise patient safety. The irony is hard to miss: systems meant to ensure accountability are creating environments where nurses feel less able to exercise their professional judgment.

What’s especially troubling is that many of these tools were introduced without meaningful input from the very people expected to use them. Nurses say they were rarely consulted during design or rollout. Training, when it happened, was often minimal — a quick video or a one-page handout — leaving staff to figure out complex interfaces mid-shift, with patients waiting. The result is a growing sense that technology is being done to them, not with them.

There’s also a deeper cultural shift at play. In an environment where every action is quantified and compared, the relational aspects of nursing — the quiet reassurance, the intuitive sense that something’s off, the time spent holding a hand — become harder to justify. If it doesn’t show up in a metric, it risks being seen as wasted time. Yet these are often the moments that matter most to patients and families, and that nurses say recharge their own sense of purpose.

Some Kaiser facilities have begun pilot programs that aim to adjust algorithms based on nurse feedback or limit surveillance during breaks. But for many, these feel like afterthoughts. What they’re asking for isn’t the removal of technology, but a seat at the table when it’s chosen, shaped, and evaluated. They want tools that reduce clutter, not add to it. Systems that alert them to real risks, not false alarms. And oversight that supports growth and learning, not one that feels like constant auditing.

The broader lesson here extends beyond one hospital system. As AI and monitoring technologies spread across industries, the experience of Kaiser nurses offers a cautionary tale. Innovation without inclusion can deepen inequities, erode morale, and ultimately undermine the very goals it claims to serve. In healthcare, where the stakes are human lives and well-being, that’s not just inefficient — it’s dangerous.

Listening to those who do the work every day isn’t just ethical. It’s practical. The best technology in the world fails if the people using it don’t trust it, understand it, or feel respected by it. For nurses on the front lines, that trust is already fraying. Repairing it won’t come from more surveillance or smarter algorithms alone. It will come from asking a simple question: what do you need to do your job well? And then actually listening to the answer.