Stop me if you’ve heard this before: An employee received a message from her boss and didn’t fully understand its meaning. The employee suspected that the message was written by AI and asked her AI tool to interpret the message. The AI responded and then asked if she wanted to send a draft response back to her boss.
The employee paused. “‘I think literally [my boss’] AI talks to my AI. “That’s the real conversation that’s happening right now,” the employee told Leena Rinne, vice president of leadership, business and coaching at Skillsoft, an edtech and skills management platform. She told Rinne: “‘I can’t crack the code of collaboration [my boss]because it’s just his AI and my AI going back and forth.’”
Rinne calls this phenomenon “social offloading,” or when interpersonal skills that require human judgment, empathy, or courage are outsourced to AI. It is similar to “cognitive offloading,” or offloading often menial tasks to technologies like AI to reduce mental effort, and has the potential to disrupt work culture.
Social offloading can look like a boss preparing for a performance review and asking the AI how to conduct the conversation. Or it could be an employee asking to write a response to a stressful email from a manager.
“If I keep asking the AI, how do I react to my boss,” said Rinne Assets, “I’m not really learning how to interact with my boss. I’m not really learning how to build a relationship with my boss.”
People are increasingly using AI in more humane ways, with the most common use being for therapy and companionship, according to a Harvard Business Review Analysis of AI usage patterns. The problem isn’t that AI doesn’t provide helpful advice, Rinne said, but rather the capabilities we lose when we rely on it too much.
“The risk then is that if AI controls emotional intelligence for us, we won’t develop these critical skills that we can use right now because we don’t know how to deal with emotional intelligence,” Rinne said.
Skillsoft uses and sells AI tools to its customers, but their tools are aimed at helping people have conversations in the real world. The CAISY product allows people to practice conversations and give feedback before having important work conversations.
Instead of, “Here’s the answer, here’s what you should say,” Rinne said, the AI instead teaches the person how to develop these intrapersonal skills. “I actually improve my ability to lead a difficult conversation or a client meeting because I have the practice.”
Table of Contents
Paying the price for cutting middle management
AI is not the cause of the problem, but rather a leadership vacuum, said Rinne. As companies flatten their organizational structures and eliminate middle management, mentoring and coaching have fallen by the wayside.
A prime example of this strategy is Meta, which has cut 25,000 jobs since 2022 and touts an AI team that has one boss for every 50 engineers. Typically, an employee-to-manager ratio of 25 to 1 is considered the extreme limit of the so-called span of control scale, but the company is fully committed to AI. With AI, some companies are reaching the limits of management.
The recent surge in younger employees appears to be a common approach also being taken by Cognizant, an IT consulting firm that says it employs more than 350,000 people worldwide and is looking to hire.
“If you can equip these people with AI, you have commercialized expertise. They have shared expertise at their fingertips. So you could have more entry-level programs and you could recruit more school graduates and get them to expertise faster,” said Ravi Kumar S, CEO of Cognizant Assets Earlier this year. While this will make the job pyramid flatter, “the asymmetry will not come from expertise, but from interdisciplinary skills,” he said.
From an organizational perspective, Rinne sees an advantage in that fewer managers can lead to faster decisions and more autonomy. However, there is still a need for managers who can translate strategy into results and action, develop talent and hold a team together, she said.
“There is a danger that organizations will start treating leadership span like a math problem, when in reality it is a skills problem,” she said.
While other generations have had decades to learn to deal with change and the organizational dynamics that come with it, “now young people are entering the workforce and are just thrown in at the deep end,” Rinne explained.
Some blame young workers’ difficulties navigating the workplace on the fact that they are generally less social. They meet and socialize less, and Tessa West, a professor of psychology at New York University whose research focuses on communication between employees and bosses, says this affects their performance at work.
“You learn a lot of skills in those early relationships that you then use in the workplace,” West said. “Negotiation is a big deal, as is compromise.”
Even romantic relationships can’t bridge the gap that Rinne sees emerging between employees and their managers. She notes that her own experiences helped prepare her for her current role as an organizational leader.
“I had great opportunities to be coached and invested in my development,” she said. “The contrast is that Generation Z is coming into the market, and I think as a digital kid you assume that they’re already ready for the pace of change or that they’re already ready to navigate.”
But leaders aren’t really equipping younger employees to handle change, communicate effectively and use good judgment, she said, reducing their competitive advantage as human-centered skills drive success in the AI age.
“We just expect them to step into this crazy whirlwind moment and be able to deal with it effectively,” she said.
A version of this story was published on Fortune.com on March 28, 2026.
More about AI in the workplace:
https://fortune.com/article/ai-communication-undermining-human-relationships-middle-management/
