AI Is Making the Workplace Empathy Crisis Worse

We expect more than ever from leaders today—including that they be strongly empathetic. But that hasn’t always been the case. Until recently, in fact, many people considered empathy in a leader to be a weakness rather than a strength. Leaders weren’t supposed to exude emotional understanding. They were supposed to be tough. In “How to Sustain Your Empathy in Difficult Times” (HBR, January-February 2024), the Stanford psychologist Jamil Zaki, a longtime student of empathy, turned that idea on its head. “Empathy is not a weakness but something of a workplace superpower,” he wrote, citing the evidence of dozens of studies. “Employees are more satisfied in their jobs, more willing to take creative risks, and more likely to help their colleagues if they work in empathic organizations. They are far less likely to report severe burnout or to develop physical symptoms of stress, and are more resilient in the face of adversity. They also tend to stay: A 2022 Gallup survey of more than 15,000 U.S. employees found that those with caring employers were far less likely than others to actively search for a new job.” The problem, Zaki wrote, is that leaders can expend so much energy being empathetic that if they’re not careful they burn out. In his article, drawing on lessons from psychology and neurology, Zaki offered leaders guidance on how to avoid that outcome by practicing what he called “sustainable empathy.” In the year and a half since Zaki’s article came out, there’s been something of a mood shift. The leaders of some big companies now boast about their successes in creating smaller workforces and mandating a return to the office, and in many quarters there’s increasing talk of how companies can best deploy empathetic AI agents to do emotional work that was once considered only possible for people. Do these trends signal a shift toward a less-empathetic leadership style, or one in which leaders will turn to AI when they need to be empathetic? To get answers to this and other questions about empathy in the age of AI, we sought Zaki out and asked him for his thoughts. . . . You wrote your sustainable-empathy article for us more than 18 months ago. What sort of response did it generate? What surprised me most was the relief readers described. It’s a gift and an honor to lead others, and people who have that power often feel it would be ungrateful and tone deaf to complain about their feelings. But leaders do feel, and the more they care about their people, the more they face the danger of burnout. Many felt seen by the article, glad to know they weren’t alone, and even happier to learn a toolkit for “sustainable empathy” during tough times. What’s changed since you wrote it? The tough times have not gotten easier! Economic turmoil, workplace churn, and the continuing rise of AI have added new stress to workplaces around the world. Uncertainty can encourage tunnel vision. Leaders focus only on what will keep the bottom line alive this quarter and think of investing in people as a luxury for better days. This is backwards. The least certain times are when it matters most to connect with people, sharing a vision and the core values that remain unchanged. Are company policies less empathetic today than they were even just last year? Few policies are always empathic or unempathic. It’s not caring to over-hire and allow people to languish in unnecessary roles. It’s not necessarily compassionate to leave workers spread out and disconnected from each other. Instead of what the policy is, leaders should focus on why and how it reflects their vision for culture. If you drag people back to the office to micromanage them while they sit on Zoom calls, of course that’s unempathic. If you bring them back for key days of creative co-work and put the value of community front and center, that can strengthen teams instead. Businessolver recently released its 2025 “State of Workplace Empathy” survey, which reported a pretty big discrepancy between the perception of organizational empathy among CEOs (who think they’re doing well) and employees (who often don’t). Can you speak a bit about that? That was Businessolver’s 10th annual survey, and it painted a darker picture than years past. This year, the number of CEOs who agreed that empathy is undervalued dropped by 28% compared to last year, signaling that they view it as more of a “nice to have” than a “need to have.” Their people disagree. Almost half of HR professionals said that leadership doesn’t support being empathic at work, and 49% of CEOs (+16 points from 2024) admit they don’t make the time to connect empathically. What about AI and its effect on perceptions of empathy? This is the question right now, but the answer is complex. Let’s start with the bad news: AI is generating an empathy gap between leaders and employees. Most workers feel apprehensive about AI, and Gen Z employees are the most likely to report that it will threaten their jobs. Like other moments of uncertainty, this great unknown calls for more empathy, not less, and more than 80% of workers say the age of AI will make human connection more important. Only 65% of managers agree, pointing to a stark disconnect that will risk workplace cultures in the years to come. Even more troubling is the rise of what I call “empathy-washing,” when companies deploy empathic-seeming AI chatbots to mask deeper issues of detachment, either between leaders and employees or companies and customers. This fools no one, and shallow algorithmic attempts at empathy (a chatbot saying “we care about your experience”) can be worse than nothing. Ironically, when workers feel starved for connection, they often turn to AI to help. In a 2023 survey, nearly half of Gen Z employees said they received better career advice from ChatGPT than from their boss. That number is almost certainly higher now. If a chatbot is running circles around you as a leader, it might be time to do some soul searching. And the good news? AI, and people’s reactions to it, reveal two crucial insights about empathy. The first comes from what chatbots do well. When humans listen to each other, we often fall for two traps: offering advice too quickly and making the conversation about ourselves instead of the other person. AI, partially because it doesn’t have a self, doesn’t make these mistakes. It follows a strict playbook: validate, express care, and only then ask if others want advice. In many cases, people find these replies more empathic than human ones. Instead of regretting this, leaders can learn from it, by trying that playbook themselves. Second, even when AIs offer skilled empathy, people want other people. When you choose to be there for someone, you sacrifice energy, bandwidth, and time, none of which you will ever get back. It is those limitations that make care beautiful. As AI continues to rise, we all have the opportunity to double down on uniquely human connection. It will become a more important skill at work, and a more precious resource for us all. Relying on AI for empathy seems like it might risk making emotional care into an automated commodity. What’s your sense about that? It’s a danger worth emphasizing. LLMs are products, not people. No matter how empathic they seem, outsourcing care to them risks devaluing it in two ways. First, unlike human connection, “LLMpathy” doesn’t demand reciprocity. Chatbots are there for you, but you don’t need to risk anything or be there for them. That makes empathy feel “easy,” but in a destructive way, by turning it into a consumer experience that can atrophy the social muscles we need to have genuine back and forth. At an institutional level, it can leave workers and customers relationally stranded. Second, AIs, like any technology, will always reflect the motives of their founders. If that means connecting people and helping them, great. But LLMs could also weaponize empathy to manipulate or addict people. Given the track record of technology companies and the tools they build, that risk seems very real.

AI Is Making the Workplace Empathy Crisis Worse

Summary.

Partner Center