AI in Healthcare: What the Next Era of Nursing Should Look Like

In July 2026, something that had long felt theoretical became real.

Montefiore Medical Center laid off 12 utilization review nurses and announced that AI software from Datavant would take over their work. These nurses reviewed patient records, confirmed that treatment was medically necessary, and helped ensure that care would be covered by insurance.

The decision unsettled the nursing community because it made one question impossible to avoid:

Will AI support nurses, or replace them?

There is no honest answer that ignores what happened at Montefiore. Although the affected roles were not bedside positions, people lost their jobs. Their concern was not simply resistance to technology. It was a response to technology being introduced without enough clarity about where people would fit.

At the same time, it would be a mistake to see AI only as a threat.

Nurses carry an administrative burden that has become difficult to defend. Research into AI and nursing shows how documentation and repetitive tasks consume time that could otherwise be spent with patients. Some studies suggest nurses spend between 25 and 40 percent of their shifts documenting care.

That is time taken away from the bedside.

It is also one reason burnout and turnover remain so high. Nurses did not enter healthcare to spend much of their day navigating forms, screens, and repetitive data entry. They entered it to care for people.

This is where AI can make a meaningful difference.

AI-powered documentation tools can reduce charting time and return part of the working day to nurses. Even 20 additional minutes can mean more time to explain a diagnosis, notice a subtle change in a patient’s condition, or support a family struggling to understand what comes next.

The value of the technology should be measured in moments like these.

AI is also helping clinical teams identify risk earlier. Predictive systems can monitor patient information, generate updated risk scores, and alert nurses when someone may be deteriorating.

But these tools do not replace judgment. They give nurses another signal to consider.

That distinction matters.

A warning generated by software has limited value without someone who understands the patient, the clinical context, and the consequences of acting too quickly or too slowly. Nurses see what data does not always capture: fear, confusion, pain, family dynamics, and the small changes that come from knowing a patient rather than simply monitoring one.

The Montefiore layoffs show why this transition needs stronger guardrails. AI may reduce administrative work in some roles. It may displace people in others.

Both realities can be true.

The Washington State Nurses Association has argued that AI should remain an innovation, not a replacement for nurses. That principle deserves to guide the industry.

Healthcare organizations must explain why these systems are being introduced, how decisions will be made, and what protections exist for employees and patients. Nurses should be involved in designing, testing, and governing the tools they are expected to use.

They understand the work better than anyone.

The future of nursing should not be framed as a contest between people and machines. The better question is whether technology can remove the work that drains nurses while protecting the work only nurses can do.

Documentation can be accelerated. Risks can be surfaced earlier. Patterns can be detected more consistently.

But compassion cannot be automated. Accountability cannot be delegated to an algorithm. And the trust between a nurse and a patient cannot be reduced to a software feature.

AI will continue to reshape healthcare.

What remains within our control is how we choose to use it.

Ravi Tandra, CEO
ProvenBase