Skip to content

Is AI Making You A Better Procrastinator?

In 1995 the Stanford philosopher John Perry wrote a short essay that he called “Structured Procrastination”. Perry’s insight was that procrastinators are rarely idle. They are often extremely busy, doing almost anything other than the one thing they should be doing: sharpening pencils, reorganising files, answering emails that could comfortably wait until Thursday. The essay eventually won him an Ig Nobel Prize in 2011, which, fittingly, arrived about 15 years late.

The trick works because it gives us two things we want at once. We get the feeling of progress, and we get to dodge the hard thinking. You could wrestle with the strategy paper, or you could clear your inbox.

The productivity writer Brian Tracy built a bestseller, Eat That Frog!, around choosing the first. Borrowing a line often attributed to Mark Twain, he argued that if you have to eat a frog, it is best to do it first thing in the morning. Swallow the hardest task early and the rest of the day feels like an achievement rather than a reprieve.

Generative AI has opened a more seductive route. Faced with a knotty problem, you can sit with it, sketch it out and risk discovering that your idea has holes. Or you can hand a half-formed thought to an AI agent and tell it to get going. It obliges instantly. Text scrolls, files appear and the token count climbs. It feels as though a great deal of work is getting done. Some of it may even be useful.

But the thinking hasn’t gone anywhere. It still has to happen, either before you press run or afterwards, when you discover at some expense that the brief was muddled. Structured procrastination has never looked so productive.

How Over-Reliance On AI Erodes Managers’ Judgement 

The cost of skipping that thinking is more than a wasted afternoon. Professor Dirk Lindebaum of the University of Bath School of Management has a name for what happens when it becomes a habit. In a paper published last month in the Academy of Management Review, written with Natarajan Balasubramanian at Ohio State Fisher, Mehreen Ashraf at Cardiff University and Patrick Haack at HEC Lausanne he describes “epistemic de-skilling”.

Managers who lean too heavily on generative AI gradually lose the practical wisdom, moral judgement and feel for context that they once built up through experience.

The mechanism is simple enough. Judgement improves with use and fades without it. Every time the tool does the thinking, the manager gets a little less practice. “Gen-AI cannot replace the lessons learned through first-hand experience,” the researchers warn. “Unlike humans, AI does not experience the world or grasp workplace complexities.”

The more hopeful finding is that the same tools can make people sharper. Lindebaum and his co-authors see the potential for “epistemic up-skilling” when AI is used to reflect on a person’s thinking rather than replace it, especially when someone has to stand behind the final decision. The difference comes down to the order of play. Think first, then ask the machine to test you, and you get better. Ask the machine first and you slowly forget how.

When Procrastination Makes You More Creative 

None of this means the answer is to start every task the moment it lands. Jihae Shin, now at the Yale School of Management, has shown that a little delay can be good for us. In research carried out at the Wisconsin School of Business with Adam Grant and published in the Academy of Management Journal, she found that people who procrastinated a moderate amount came up with more creative ideas than those who started straight away. Those who left things to the last minute did worst of all.

The delay helped because people had already engaged with the problem. While their attention drifted elsewhere, they restructured the task in their heads and drew on knowledge they hadn’t thought to use at the start. The benefit also showed up only among people who cared about the work or were expected to be creative. For everyone else, putting things off simply made the work worse.

That helpful for anyone working with AI. Wrestle with the problem first, then let the agent run while your mind wanders, and the wandering may do you good. Hand over a half-formed idea instead of thinking, and the daydreaming is just procrastination with the meter running.

Why Business School Teaches You To Beat Procrastination 

Business schools have their own AI policies to argue over. But the way an MBA is taught happens to be a very good training ground for doing the uncomfortable thinking first.

Start with the case method, where no agent can take a cold call for you. When a professor turns to you in a room of 90 classmates and asks what the CEO should do, you need a position, the reasons behind it and a sense of where it might crack. Students who skimmed a summary are found out within a minute. Those who wrestled with the case the night before, and then slept on it, tend to have the more interesting answer. Jihae Shin would not be surprised.

Then there is the sheer volume. An MBA hands students more reading, recruiting events, club commitments and group projects than anyone could possibly finish. Tidying the desk is not an option, and students learn quickly to decide what really matters and do the hard part of it first, which is exactly the skill that structured procrastination erodes.

Study groups add accountability, the ingredient Lindebaum identifies as the difference between AI making people sharper or duller. A teammate who arrives with a confident but hollow analysis is caught out in front of people whose respect they want to keep.

And underneath it all, business school teaches delegation. Leaders spend their careers handing work to others, and the good ones learn that a vague brief produces vague work, however talented the team. An AI agent is simply the most obedient team member a manager will ever have. It will run with a bad brief all night and never complain.

One Question To Ask Before You Delegate To AI

Perry’s advice to fellow procrastinators was to put tasks at the top of the list that look urgent but aren’t, so that the avoidance itself gets something done. It was a joke with a serious point. In the age of AI the stakes of that joke have changed, because the busywork now looks exactly like the real work.

So here is a simple test before handing anything to an agent. Could you explain to a sceptical colleague what you want, why it matters and what a good answer would look like? If you can’t, the machine can’t either. It will just take longer, and cost more, to tell you so.

Interested in this topic? You might also like…

Leave a Reply