The AI Hangover

A joint study between Microsoft Research and Bocconi University tracked over 3,500 knowledge workers across multiple organisations. The findings during AI use were predictable: higher output quality, faster task completion, increased engagement. The expected returns on investment.

The finding nobody planned for came after. Workers who had AI tools removed reported an 11% drop in motivation and a 20% increase in boredom, both measured against their own pre-AI baselines. Not compared to the AI-assisted period. Compared to before they ever used AI at all.

They didn't return to normal. They went below it.

This pattern should be familiar to anyone who studies tool adoption. Once you've experienced a better way of working, the old way feels worse than it did originally. It's the same reason going back to a flip phone feels intolerable, even though you used one happily for years.

But there's a sharper edge here. Organisations are rolling out AI tools with the assumption that adoption is the hard part. Get people using it. Measure the productivity gain. Declare victory.

Nobody is planning for the withdrawal.

Three things leaders should be thinking about.

Don't pilot and pull. The standard approach of "let's try AI with one team for three months and then evaluate" has a hidden cost. If you give a team AI tools and then remove them, you've made that team's experience of their job measurably worse. Pilot carefully. If you're going to trial, be upfront about what happens after.

Budget for continuity. AI tool costs are ongoing. If your organisation adopts AI tools and employees build workflows around them, removing those tools to save money has a productivity cost that exceeds the subscription fee. Factor retention into the business case from day one.

Watch the psychological signals. Engagement surveys, team energy, frustration levels. If AI tools change or degrade, monitor the human response. The 20% boredom increase is a leading indicator for disengagement and attrition.

If AI makes work feel better, and removing AI makes work feel worse than before, what does that say about the work itself?

Maybe the tasks that AI automates were always a drag on motivation. Maybe the boredom was always there, masked by the busyness of doing it manually. AI didn't create a dependency. It exposed how much of knowledge work was tolerable only because there was no alternative.

That reframing doesn't make the management challenge easier. But it does make it more honest.