Nobody Saw It Coming: Why Everyone’s More Stressed Since They Got AI

Jane, a marketing manager for a fast-growing software engineering firm, estimates that she saves about 8-10 hours a week synthesizing and summarizing survey data, thanks to her AI assistant.

The results almost always have to be tweaked, she admits, and sometimes they’ve been all wrong, requiring rework. But still, she’s pretty happy with her newfound productivity gains.

When her manager asks how she plans to use the extra time, Jane says she’ll be creating a series of detailed recommendations based on survey data summaries, and she’ll start running surveys and focus groups for all planned products, rather than just a few deemed as high priority. And if she has a chance, she’ll show her teammates how AI might save time for them.

Jane is jazzed about the prospect of being able to do more things, faster, with help from AI. But at what cost?

Productivity gains from the use of AI may sound great, but they can lead to an intensification of work, longer hours and more pressure. This edition, based on a recent Harvard Business Review article, explores how and why.

Task expansion: It may be tempting to want to take on new or expanded tasks when AI can do some of the heavy lifting. In Jane’s case, she’s taking on more of the same tasks she’s already been doing, and she’s initiated a brand new task that she’d never had time to do before. She might also offer to take over tasks from overburdened colleagues. So what’s the problem?

Jane’s cognitive load will be heavier, and the time she’s saved with AI will likely be outweighed by the time she’ll be spending on new or expanded tasks. As far as taking on teammates’ tasks, they’ll have to take time to review her work and quite likely revise it, an unplanned time-sink.

Blurred boundaries: Using Gen AI is fun! It’s easy! It’s always waiting for you! You can use it on any device! It makes you feel smarter! No wonder so many people are using it for all sorts of things wherever they are. What’s the harm?

Well, for one thing, badly needed downtime may even be harder to get, as more people befriend their AI bots and forsake the people around them.

One recent example: A group of friends and I were chattering away about the latest book we’d read. Someone raised a question about the author. Everyone whipped out their phones to prompt their AI assistant for the answer. Suddenly what was a lively conversation turned into several minutes of silence as people scrolled their phones.

Bottom line: Since your favorite AI assistant can give you an answer about almost anything, human interactions may become a lot less interesting or even necessary for many people.

Multitasking: Some people are adept at task-switching. But until someone finds a way to cleave the brain in two, humans can’t perform multiple thinking tasks at one time. Ah, but this is where AI can help! Or can it?

Yes, we can create prompts in multiple chat threads at once, but once the responses start pouring in, we actually have to attend to those responses. This may require running the query multiple times with different prompts, validating the responses against other data, reviewing the results with a colleague or manager, selecting which responses to paste, etc.

In short, our task-switching goes into overdrive, diminishing our ability to think critically or exercise sound judgment. It’s no wonder we’re seeing so much AI-generated “workslop,” which sacrifices quality thinking and writing for speed.

Practical Actions for Leaders:

As a team, create principles and norms governing the use of AI for team tasks, especially those that are interconnected or handled by multiple people.
Agree on the highest and best use of AI for your team at least initially, and assign accountabilities for automating, expanding or outsourcing certain tasks through the use of AI.
Make it a routine part of your team meetings to ask people to share what they’re using AI for, what they’re learning, and how the use of AI is affecting their scope of work and the work of the team overall.
As people experiment with AI to get more of their work done, pay attention to “voluntary overload.” Notice when people are taking on more work as a result. Find out how this expanded work is affecting how long they work, their stress levels and their level of engagement. This AI honeymoon phase can’t last forever, so check in frequently.
Know that time saved from AI does not necessarily mean reduced pressure. In fact, the opposite may be true, especially those who are using AI for multiple tasks at once.

While some (though certainly not most) organizations have established clear governance structures and generally accepted use cases for AI, it’s the job of leaders to keep a watchful eye on the impact of AI use across their teams. To what degree is work intensifying, and for whom? What interventions may be needed, and when?

Short-term productivity gains may be exciting at first, but they can mask long-term unsustainability.