The Speed Mismatch: Why AI Adoption and Employee Well-Being Are Moving in Opposite Directions
Here’s a stat that should stop every HR leader mid-scroll: 77% of workers say generative AI has increased their cognitive workload, not reduced it. Meanwhile, 92% of companies plan to keep investing in AI over the next three years. We’re pouring resources into tools that are supposed to make work easier, and the people using them are telling us, in large numbers, that it isn’t working that way.
That gap is what I’ve started calling the speed mismatch. Technology changes at the speed of technology deployment, but humans adapt at the speed of learning, trust, and meaning-making. While most organizations are measuring the success of AI rollouts by adoption rates and productivity gains, they’re losing sight of how it’s impacting their workers’ well-being. In general, we know that employee well-being has significant impacts on organizations and their bottom lines. Consider, for example, that ComPsych data shows mental health leaves increased by 300% from 2019 to 2024, or that one in three employees say they’re likely to quit within the next six months because of burnout.
If HR doesn’t own this conversation, no one will.
Four Places AI Quietly Reshapes How We Work
When I talk to HR leaders about this, I break it into four areas, because “AI and well-being” is too abstract to act on. Each one has a specific risk and a specific fix.
- Think. AI is changing the shape of cognitive work, not just the amount of it. A manager who used to make 20 decisions a day might now review 60 AI-generated recommendations instead. That’s not less thinking — it’s a new kind of mental load, on top of the old one. The risks here include automation complacency (trusting the output because it sounds confident), AI sycophancy (the tool agreeing with you so readily that it quietly reinforces your existing biases), and constant context-switching that fractures attention. The fix isn’t complicated: write your own first draft before asking AI to polish it, ask for two or three competing perspectives instead of one answer, and treat “get an answer” and “get a good question” as two different goals.
- Feel. Resilience is the ability to stay calm and functional when things are hard. You build it the same way you build any skill: by practicing it under real stress. The concern with AI is that it’s available 24/7 as a source of comfort, and comfort that’s always available can crowd out the harder, more valuable work of developing your own coping skills. Used well, AI can help you find information and resources, but it’s a poor substitute for a person, and it was never designed to be one. Make sure you are pushing yourself for those interactions that matter, avoiding emotional overreliance on AI, and continuing to flex and build your resilience muscles.
- Connect. This is the one I think gets underestimated. One in five employees already report feeling lonely at work, and AI has a way of quietly replacing the small human interactions — the quick question to a colleague, the “does this make sense to you?” — that hold teams together and forms powerful connections. “Social snacking” can occur making AI feel like connection, but it doesn’t build the same meaningful relationships. Left unchecked, it can degrade social skills and reset our expectations of how easy human interaction “should” feel. Simple counters help, such as committing to one AI-free, human-intensive day a week, and defaulting to asking AI to point you toward a person rather than answering the question itself.
- Thrive. AI is a machine. It never gets tired, never needs a break, and it will never be the one to tell you to stop. That’s precisely the problem for physical well-being. AI removes a lot of the natural interruptions that used to break up a sedentary day, such as walking to a colleague’s desk, discussing an idea while grabbing coffee, or a hallway conversation. AI is designed to keep you engaged, not to protect your body. Setting a hard stop time before you start an AI session and treating usage limits as a built-in break rather than an annoyance, are small habits that matter more than people expect.
The HR Opportunity: From Enablement to Resilience Champion
Most AI change management stops at training people to use the tools. That’s necessary, but it’s not sufficient, and it leaves HR in a purely operational role. There’s a bigger opportunity: HR can be the one asking whether the way we’re deploying these tools makes people feel more capable and more connected, or less.
Practically, that means a few shifts. Redesign the work, not just the workflow — protect deep work time, and create AI-free collaboration spaces on purpose rather than by accident. Pair every AI literacy investment with a real conversation about what people are worried about; psychological safety is the actual prerequisite for adoption, not an afterthought. And change what you measure: track well-being indicators like perceived relevance, psychological safety, and social connection right alongside your productivity dashboards, because the two are more connected than most rollout plans assume.
None of this requires a massive program to start. This week, you could block 30 minutes with a few teams to ask directly: what’s the plan for the human experience in this change? You could add a short human debrief that just includes a conversation after your next AI training session. You could run a managers-only session first, since leaders set the tone for how a team experiences any change. And you could add one well-being metric to whatever you’re already tracking for adoption and productivity.
Where to Start
The productivity narrative around AI has outrun the people narrative, and closing that gap is HR’s job now, whether or not it was written into the AI rollout plan. The good news is that the four areas above — thinking, feeling, connecting, and thriving — are all addressable with fairly simple, concrete habits. None of it requires employees to use AI less. It just requires using it more intentionally.
To make that easier for your own workforce, ComPsych put together a short, practical resource called Well-being in the Age of AI: Four Paths to Healthy Use, developed by Jennifer Birdsall, Ph.D., ComPsych’s Chief Clinical Officer, and Alexandra Samuel, Ph.D., an AI researcher and author. It walks through each of these four areas in plain language, with concrete steps people can start using right away. If you’re looking for a low-lift way to get ahead of the conversation your organization probably isn’t having yet, that’s a good place to start — share it with your team this week, or read it yourself first if you want to see what your people are going about to ask you about.
