Nurses say technology is making their jobs and patient care worse

Seven current and former nurses told The Markup that those who spend more than 15 minutes on a call with a patient routinely face criticism from Kaiser management or get called into performance evaluation meetings. Call time, they said, factors into monthly performance scores they receive.

In addition to tracking call length, they said Kaiser uses software that tries to predict on a daily basis whether they’re being unproductive or failing to answer calls quickly. Artificial intelligence systems have also been used to rate their empathy and tone of voice.

Their comments come as the California Nurses Association begins negotiating a new contract with Kaiser this month with AI a likely issue. Kaiser nurses went on strike against AI for one day in March and picketed against AI last fall. The CNA is bargaining for 25,000 nurses, including 1,000 in call centers. 

At the same time, California lawmakers are considering several bills regulating AI in the workplace, including one that would protect from retaliation doctors and nurses who override automated care recommendations.

Kaiser defended its use of AI, saying it deploys the technology with human oversight and with patient safety in mind. It said it does not use “average handle time” to assess performance.

Kaiser Permanente is the largest private employer in California, providing healthcare services to more than 9 million people in the state and to 3 million other Americans. That means the company’s use of artificial intelligence could set important precedents for managing workers with AI. It could also have a big impact on patient care, providing an early example of how the healthcare sector balances cost-cutting automation with human presence or touch.

Raquel Alvarez Sanchez, a Kaiser Permanente advice nurse in Vallejo since 2010, said she was on a call with a patient who was suicidal last year that took more than an hour because she had to wait for police to arrive before hanging up. She tried to make the man feel cared for, even though she was cognizant that staying on the call that long would throw off her average call time for weeks and could lead to questions from management. Sanchez, a union steward, said she’s accompanied colleagues to performance evaluation meetings, where they were found to have done everything right on a call — except staying on the line for more than 15 minutes. She said she hasn’t seen nurses get fired for doing that, but she fears that continued pressure can lead nurses to quit or retire early.

“I think at some point all of the nurses have been talked to about their average handle time,” she said. “The only thing I can think of is they’re doing it for profit.”

Another nurse who spoke with The Markup on condition of anonymity due to fear of retribution described how that surveillance affected a call with a patient last year. Initially, she thought her patient, an elderly woman who just received a terminal cancer diagnosis, was suicidal, but quickly came to understand that she was in shock and really needed somebody to talk to.

The nurse wanted to take time to show compassion or comfort to the woman, who acts as a caretaker for her daughter, but she stopped herself out of fear it would hurt her monthly performance score and lead to a reprimand from her manager. She became a nurse to provide people with compassionate care, but “I had to ask myself: Am I going to get disciplined for going off script or saying more than what is necessary?”

Kaiser Permanente says its performance evaluations help improve patient outcomes. A company spokesperson said, “Kaiser Permanente does not use Average Handle Time to assess agent performance or enforce call time metrics. Any tools used in contact center settings support our quality assurance efforts and have human review and oversight.” In a statement provided to The Markup and CalMatters, spokesperson Vincent Staupe added that Kaiser uses AI responsibly, with human oversight, and by “prioritizing patient safety, privacy, and equity,” but he said, “As a large organization, we do not share specific information about internal technology systems for security and operational reasons.”

Is technology putting patients at risk?

It’s not clear how patient care is affected by algorithmic management, nor is the impact of limiting the length of triage and advice calls on patients. Kaiser call center nurses can’t say for certain whether the pressures they face result in adverse outcomes for patients because their contact with patients ends after they hang up the phone. A 2024 public records request by CalMatters to the California Department of Managed Health Care found no complaints by patients against Kaiser related to call times. But nurses insist the risk to patient safety and quality of care is real. 

Consumer Watchdog patient advocate Michele Ramos said many Kaiser patients begin their care on the advice line. They later complain to her, mostly about things that happen in Kaiser facilities, but “I can see now where a lot of the problems” start, given the call constraints nurses are under. 

Ramos said the time pressures may fit a broader pattern at Kaiser of putting costs over quality. The health giant was hit with a record fine, $50 million, as part of a settlement over findings from the California Department of Managed Health Care that it delayed behavioral health appointments beyond statutory limits and too often moved patients into group rather than individual therapy. Kaiser also settled with the U.S. Department of Labor after investigations into its substance use and mental health services. Kaiser faced criticism in 2002 for paying bonuses to call center workers who aren’t nurses for keeping calls short, though call center nurses who spoke with The Markup and CalMatters said they encountered no such practices today.

“Kaiser’s been known through the years to manage dollars over managing care, and I think this would be a contributor to that, which is only going to fail patients,” Ramos added.

Nurses said they are pressured to stay under 15 minutes even for the sorts of calls that often take more time, like diagnosing a patient with multiple symptoms, chronic illnesses, new parents in need of advice and assurance, people who desire extended health education, or people who are overwhelmed after receiving life-altering news who could use some compassion. Nurses say calls that involve interpreters often take 30 minutes or more. About four in 10 Californians speak a language other than English and half of them do not speak English well, according to a state environmental health agency.

“The amount of time that Kaiser is giving us to complete a call is sometimes not safe,” said one nurse, who asked to remain anonymous due to fear of retaliation.

“People can get hurt,” said Charlotte Capulong, who has worked in nurse call centers for 22 years and helped organize Kaiser nurses against the AI tone-of-voice tool. Capulong said nurses felt harassed by managers in meetings she attended as a union rep, even if they successfully carried out all other duties of their jobs except completing calls within 15 minutes.

“You aren’t calling Comcast. We’re dealing with life here,” she said.

Nurses are instructed to stick to a script on phone calls and give no more than two to three pieces of advice, Capulong and other nurses said, which means they may sometimes need to decide whether to withhold advice or face a performance evaluation hearing.

The nurses say artificial intelligence could make the surveillance nurses encounter on the job worse.

In summer 2024, Kaiser began testing an AI tool that attempts to assess empathy and tone in the voices of nurses and patients, according to nurses who spoke with The Markup and CalMatters. In response, nurses circulated and signed a petition in favor of the right to patient privacy, more transparency,  and the right to exercise their professional judgement and encouraged management to involve nurse’s input and feedback. The signature campaign used the same tag line that nurses used at protests outside San Francisco hospitals earlier that year: “Trust nurses, not AI. The AI tests ended in November 2024, but union representatives were told that managers may bring the program back in the future. 

Nurses reported feeling harassed by existing surveillance, “and that was intensified when they said we’re going to use AI to evaluate our calls and grade us,” said Sanchez.

Another nurse speaking on condition of anonymity said “AI did not understand our job and would grade us wrong all the time.”

A Kaiser spokesperson declined to respond to questions about the AI tool or answer questions about the use of AI and other automated systems in the company’s call centers and healthcare facilities, including for evaluating nurse performance or whether patients were informed about the use of AI to evaluate their empathy and tone.

Nurses also said they get little time between calls, even if that call involves speaking with a patient who is suicidal, experiencing a mental health episode, or near death. In years past, nurses got around 10 minutes to finish writing notes in a patient’s chart or collect themselves after a particularly tough call. Today they say they typically get 30 seconds or less when lines are busy, although more at slow times, like late at night, or if they get a manager’s permission after a particularly challenging call. The overall pace, they say, can lead to mistakes like missing important cues into a patient’s wellbeing.

CNA reps declined to talk about specific provisions they intend to seek related to AI ahead of their talks with Kaiser this summer.

How surveillance and AI shape nursing

Critics say excessive workplace monitoring can lead to lower morale as employees feel less trusted and autonomous, relegated to being no more than algorithm monitors. UC Berkeley Labor Center Technology and Work Program director Annette Bernhardt has warned that algorithmic management can turn people into fleshy robots, echoing complaints from an Amazon factory worker who The Markup and CalMatters interviewed last year. A 2023 academic survey of call centers in four developed countries found that using AI for management or monitoring left workers with less time between calls and more likely to feel emotionally drained by their work. Nearly half of respondents said that AI tools made their jobs more stressful. A prior study by the same researchers, Virginia Dolleghast of Cornell University and Sean O’Brady of McMaster University found that performance monitoring leads to higher rates of emotional exhaustion.

Dolleghast, who has studied the impact of surveillance technology on call center workers for more than a decade, said what Kaiser call center nurses are experiencing is part of a broader trend: Across different industries, persistent surveillance is increasing stress levels for workers who are resolving complex, emotionally-charged issues. 

“Stress and burnout can lead to more mistakes across a range of areas, and in the healthcare setting that is much higher risk because you’re dealing with people’s lives and their health,” she said.

The converse can be true: Workers who are given more discretion to decide the pace and timing of their work experience higher levels of job satisfaction and less absenteeism.

Nurses nationwide are more frequently encountering artificial intelligence and similar software systems in the workplace. Half of more than 2,000 nurses who responded to a 2024 survey by the National Nurses United union said their employer uses algorithmic systems to analyze health records. Such systems can do things like determine how fragile a patient is or predict how many hours of care they will need. Two-thirds of the surveyed nurses said their own assessments had at some point disagreed with a computer-generated prediction. Six out of 10 respondents said they don’t trust their employer to prioritize patient safety when using AI.

Pa Vue has worked as a nurse in call centers for the better part of the past decade. She said she and other Kaiser nurses routinely have conversations with managers about call efficiency and receive evaluation scores once a month. She recalls having a score reduced for repeating advice to a patient that she worried had unusual symptoms and possible heart issues.

As a union representative in some performance meetings, Vue has seen managers raise efficiency questions about calls they deem too long. She’s also seen nurses receive lower performance scores if they go against software recommendations based on their professional opinion or make an appointment for a patient without consulting a doctor.

She believes that efficiency aims accelerated by technology can hinder a nurse’s ability to focus and reduce the quality of care that patients pay for.

“I’m not against the use of AI as long as it’s beneficial to the patient but in this particular use [empathy and tone monitoring] it’s to increase productivity and improve efficiency and cut costs. Kaiser is forgetting we aren’t just a call center for customer support, we’re nurses, and we’re there to take care of patients,” she said.

As AI improves and businesses push workers to use it, unions are, in turn, increasingly demanding that employers address issues raised by AI when bargaining for new contracts. Surveillance technology has become a common way for managers to collect data about workers in a number of industries, used for everything from improving safety to hunting for ways to increase profit gains or train AI to do a job.

At Kaiser, AI is a key issue not only among nurses but also for mental health workers, 2,400 of whom are in contract negotiations in Northern California with Kaiser Permanente. Kaiser therapists have said they are concerned about use of therapy session transcripts to train AI models and about the health-care giant using AI to take their jobs. National Union of Healthcare Workers spokesperson Matt Artz told The Markup that contract negotiations are ongoing.

How Kaiser uses AI

Kaiser Permanente is exploring or using AI in many parts of the healthcare experience far beyond nurse call centers. Kaiser uses AI to identify patients in hospitals at risk of adverse events by evaluating data on their electronic health records. An AI system called Preventus is used to determine when to discharge patients. Doctors and therapists use Abridge to record interactions and translate speech to text during in-person visits with patients instead of taking notes. Remote monitoring with AI for patients that need extra care has been tested at Kaiser Permanente facilities in the Bay Area, according to nurses who encountered the technology in the course of doing their jobs.

National Nurses United and CNA President Cathy Kennedy sees the use of AI to detect nurse empathy as part of a long series of steps by Kaiser to limit their autonomy and make them more efficient. She believes AI threatens to automate and fragment the work that nurses do, and companies developing and deploying AI systems should establish that those systems are effective and equitable before deploying them.

Notification of new tech deployments is part of the nurse union’s contract with Kaiser but sometimes nurses don’t receive notification, CNA says. So union leaders are attempting to track the number of AI models in use at Kaiser Permanente, advising its members to inform them when they encounter new tech. This paves the way for CNA to push back as it did with the empathy and tone AI last summer or as it did when it stopped a pilot program that would have replaced nurses that sit at the bedside of confused patients with cameras.

Debru Carthan, a Kaiser radiologist, is on the front line of worker-management fights over AI at the company. A member of Service Employees International Union, she is also part of the Coalition of Kaiser Permanente Unions, where she sits on a committee to discuss use of AI and emerging technology at Kaiser. The coalition also has a “see something, say something,” campaign for frontline workers to report when they notice AI deployments; the coalition says that too often management quietly implements AI into workflows without notice or worker input. She worries that the AI tone detector used on advice nurses could discriminate against nurses from different cultures and has come to believe that the use of AI in healthcare generally has more to do with money and corporate greed than patient care.

California lawmakers have responded to worker AI concerns both inside and outside the healthcare sector. They tried and failed last year to address how AI impacts workers like call center nurses. Assembly Bill 1018 and Senate Bill 7, two bills endorsed by the CNA, would have required employers to inform workers before using automated systems on the job to do things like promote or discipline workers or evaluate job performance, but Gov. Gavin Newsom vetoed SB 7, and, facing strong opposition from companies including Kaiser Permanente, AB 1018 failed to pass for the third consecutive year. 

Earlier this year, lawmakers reintroduced a new version of Senate Bill 7, now called Senate Bill 947. Another bill would prohibit employers using AI to predict the emotional state of their employees. Yet another bill would protect doctors and nurses from retaliation if they override recommendations generated by an automated system and require healthcare providers to supply employees with an inventory of automated systems once a year. Kaiser declined to share a list of AI systems in use when asked by The Markup and CalMatters.

Altogether, CNA and the affiliated California Labor Federation support roughly half a dozen bills to regulate the use of AI in the workplace. Calling AI a central issue in the next presidential election, members of the California Labor Federation and labor leaders from Democratic primary states held a press conference in Sacramento earlier this year to say that if Newsom wants to become president then he needs to pass laws protecting workers from AI. “It’s an ongoing fight, and it’s a fight well worth having,” Kennedy said. “Whenever there are other unions in discussion about artificial intelligence we are in solidarity with them.”

The nurse that withheld compassion to a terminal cancer patient she thought was suicidal said she believes monitoring and scoring systems turn nurses into automatons that check boxes.

“I used to use humor as a way to help patients heal, and I don’t feel comfortable doing that here because I know the calls are being recorded. You can always tell when a patient appreciates the humor or your personal compassion, but I don’t feel like call centers have tolerance for that because that’s not part of the script,” she said. “That really takes away from the whole point of being a nurse and what patients come to know from nurses.”

This story was reported with contributions from Lam Thuy Vo and Ana Ibarra.

This story was produced by The Markup and reviewed and distributed by Stacker.