I am afraid we are using AI for things that don’t need to be done.
We are making non-value-added work faster, which means we can do more of it. But doing waste faster doesn’t make it less wasteful.
Ignoring the controversy around Elon Musk, there is an interesting lesson from Tesla’s early Model 3 production.
Tesla tried to automate large parts of the manufacturing line. Eventually, Musk admitted they had gone too far. Some of the automation actually made production more complicated.
At Tesla’s 2018 shareholder meeting, Musk talked about one of their mistakes: trying to automate tasks that were easy for a person but difficult for a robot. His takeaway was essentially to start with people and then automate the things that are painful or difficult for people to do.
I wonder if we are about to make the same mistake with AI.
Right now, one of the things AI is especially good at is taking mundane tasks and removing some of the toil from our jobs. That sounds great. The part I am afraid of is what happens next.
Do we actually remove the toil?
Or do we just create more of it because it is now cheaper and faster to produce?
There is already some evidence that this distinction matters.
A 2026 review from the International Labour Organization looked at research on generative AI and productivity. It found real productivity gains from AI, but those gains were uneven. More interesting to me, the time workers reported saving with AI had not necessarily translated into corresponding increases in organizational output.
In other words, making a task faster isn’t automatically the same as making the organization more productive.
There is also research asking whether AI reduces workload or simply intensifies it. A 2026 systematic review of research on generative AI and academic workloads found evidence for both. AI can save time on individual tasks, but it can also create new work around checking, revising, learning, and managing AI-generated output.
What are we doing with the time AI saves us?
If AI makes it easy to create another report, another summary, another presentation, another status update, or another document, we may end up with more things for someone else to read, review, respond to, and manage.
Maybe managers should look at AI a little differently.
Every time we find a task that is a great candidate for AI, maybe our first question shouldn’t be, “How can we automate this?”
Maybe it should be:
“Why are we doing this at all?”
AI might be more than a tool for eliminating toil.
It might be an indicator of work that should end.