Perseverance And Grit Are At Risk
Picture yourself stuck on a hard problem. The kind that used to make you sit there, jaw tight, turning it over until it cracked. Now the help is one tab away. So you reach for it…a little faster every time. A little earlier every time. Until one day the stuck feeling shows up and you do not even try. You just ask.
Researchers at Carnegie Mellon, Oxford, and MIT put numbers on exactly that. In a randomized study, they gave half a group an AI assistant and set everyone loose on hard problems. The AI group leaned on the tool, no surprise. But, the group with AI quit sooner. They abandoned the hard problems faster. And when the tool was taken away and they had to work alone, they performed worse than the people who never had it.
Sit with that. The damage was not only that they got worse at the task without the crutch (skill). They also quietly lost the will to do the hard work. They lost both skill and will. The grit drained out.
There are many names for the thing that drained. Angela Duckworth famously titled her book about the subject “Grit”. The will to stay with something hard long after it stops being fun. She argued it can matter as much as talent.
And this new study suggests AI quietly taxes it, one easy exit at a time.
For a year the warning has been about skill. Use it or lose it. This one is sharper.
In addition to the skill, you lose your perseverance, your grit, your “sticktuitiveness”. And the fight is exactly where the skill was built.
Grit Breakdown Mechanism
A skill is not built in the moment you get the answer. It is built in the moment right before, the part where you are stuck and you stay. Robert Bjork, a memory researcher at UCLA, names the good version of this “Desirable Difficulty”. What that means in plain language is the strain is not the obstacle to learning. The strain is the learning. The struggle is the workout.
Put it even more simply. You do not grow in your comfort zone. And conversely, you are not comfortable in your learning zone. The growth lives in that uncomfortable stretch just past what you can already do. The edge is the whole point. And the edge is exactly what AI offers to take off your hands.
AI does not just hand you the answer. It removes the workout that comes right before the answer. And because the workout felt bad, losing it feels like a gift. You feel faster. You feel smoother.
Nothing feels broken. But the muscle that pushes through difficulty quietly weakens, so the next time you are stuck you have a little less in the tank to stay there. So you reach for help a little sooner. Which removes the next workout. Which weakens it again. That is the trap. It is a loop.
The people most exposed to AI are the ones doing the hardest, most valuable thinking, because they have the most struggle to offload. Your edge was never just what you know. It was your tolerance for staying in the hard part longer than everyone else. That tolerance is the thing draining, and it drains quietest in the people who used to have the most of it.
Consider a butterfly struggling mightily to get out of its chrysalis. Without that struggle its wings are small and shriveled, and it never flies. It spends its short life crawling on the ground. Because the squeeze through that tiny opening was never the obstacle to flight. It was the mechanism to prepare its wings for flight. Remove the struggle and you do not free the butterfly. You ground it.
So do not avoid the struggle. Protect the beneficial part of the struggle. Three moves to protect your edge. Let me show you each one.
Move 1 - Protect the First Ten Minutes
The answer is one click away, in your favorite AI tool. The benefit is in the ten minutes before you open it.
On any problem that matters, give it ten honest minutes of your own fight before you open the tool. Not the whole job. Just the first real wrestle, because that wrestle is the rep that keeps you strong.
Make it physical. Set a timer, then get out a notebook and pen, pencil and paper, or stand at a whiteboard. Sketch what you already know. Draw the diagram. Write down the constraints. Do I have a hypothesis? What does my experience tell me that the AI might not know?
The work on that page is the rep that builds the mind and the will. It leads to a sharper prompt, and, it leads to better AI output. You stop showing up empty, asking it to think for you. You show up with a hypothesis and the constraints already in hand, so the tool builds on your thinking instead of replacing it. You stay the one leading.
Think of it like a workout. You do not skip the sets and still get stronger. Skip straight to the answer and you got the result without the one rep that was building you.
The tool will still be there at minute eleven. You are not using AI less, you are making sure the struggle lands before the answer.
The failure mode is reaching for the easy button the second it gets uncomfortable. Discomfort means you are in the rep, not that you should bail. Ten minutes. Timer on. Tool closed.
Pay the ten minute tax, and the skill stays yours.
Move 2 - Protect Your One Hard Thing
Offload almost everything. Guard the one thing your reputation is built on.
You do not have to struggle through everything. That is not the move, and it would be exhausting. The move is to pick the kinds of problems where the struggle is the entire point. The thinking that is your reputation. The thing people come to you for.
That one thing is not random. It usually sits at a single intersection: what your situation actually needs, the way you are wired, and the new tools now on the table. That overlap is your secret weapon, and it is the last place you hand the struggle away. I made a whole piece on finding that intersection, in the previous article.
This is the lift you shouldn’t outsource. Everywhere else, offload freely, let the tool carry the load, save the hours. But on that one, you do not get to reach early. There, you stay in the discomfort on purpose, because that is the muscle you refuse to let drain.
Name it out loud. Write it down. If you cannot say what your one hard thing is, that is the first work, because a muscle you have not named is a muscle you will offload without noticing. The expert hits a wall and the tool is right there, so the expert reaches. Naming the one thing is how you decide, in advance, where you will not.
Guard the one rep that is yours, and let the tool carry the rest.
Move 3 - Protect the Struggle Inside the Tool
You can let the tool lift the weight, or you can let it spot you. Only one builds muscle.
Here is the move for the other twenty-three hours, the ones where you are going to use AI no matter what I say. Good. Use it. The question was never whether you open the tool. It is what you ask it to do once you are there.
There are two ways to use AI on a hard problem. You can ask it to hand you the answer, or you can ask it to make you earn one. The first is a forklift. The second is a sparring partner. Same tool, opposite effect on you. A forklift carries the load so your muscles never fire. A sparring partner hits back just hard enough that you come out faster and sharper every round.
So change the ask. Instead of “give me the answer,” try “ask me three questions before you answer.” Instead of “write this for me,” try “here is my draft, attack it.” Instead of taking the first response, make it hold the solution back until you have taken your swing. Tell it to play the skeptic, the examiner, the coach who will not let you coast. You are still using AI on every problem. You are just refusing to let it do the one part that was building you.
The failure mode is the silent default. Left alone, the tool optimizes for giving you the answer fast, because fast looks helpful. Helpful and good for you are not the same thing. If you never change the instruction, the easy version wins every time, and the easy version is the one that drains you.
Let the tool spot you, not lift for you. Same tool. Opposite outcome.
Your One Move
So here is where we land. AI removes the struggle right before the work, and the struggle was the thing building you. Lose that quietly enough times and you do not just lose a skill. You lose the grit, the will, the perseverance that is needed to live life well.
Years ago I had my kids read a book called Do Hard Things. The whole spirit of it was a rebellion against low expectations. Choose difficulty on purpose, because that is where you grow. That idea is more necessary now than the day it was written, because the easy exit has never been one tab closer.
And the discipline does not stay in one lane. Push harder on your last set of pushups and you push longer on a hard problem. Hold your focus through a brutal interval and you hold it through a brutal spreadsheet. Researchers put people through a two month exercise program and watched their self control climb in places they never trained, less impulsive spending, better study habits, sharper focus. Discipline is not a tank you drain. It is a capacity that compounds. Every hard thing you choose makes the next hard thing easier to choose.
I try to choose the harder path in small ways, too. I used to look for great parking places close to my destination. Now, I often choose to park far away. Why do I do it? So that I can get in the extra steps. The key is to get the steps and reps in every day to stay strong.
So here is the one move, and you can run it this week. Pick one hard problem that matters to you. Before you open AI, give it ten honest minutes of your own fight. Tool closed. Get something on paper. Just the first wrestle.
That ten minutes is the rep that keeps your grit yours. Do it once and you will feel the difference. Do it every week and the drain never gets a foothold.
Subscribe for weekly ideas on turning research into practical advantage.
AI will gladly take the struggle off your hands.
Decide, on purpose, where you keep it. That decision is the whole game.
Sources
The research behind this piece:
1. Liu, G., Christian, B., Dumbalska, T., Bakker, M. A., & Dubey, R. (2026). AI Assistance Reduces Persistence and Hurts Independent Performance. arXiv:2604.04721. https://arxiv.org/abs/2604.04721
2. Duckworth, A. (2016). Grit: The Power of Passion and Perseverance. Scribner.
3. Bjork, R. A., & Bjork, E. L. (2011). Making things hard on yourself, but in a good way: Creating desirable difficulties to enhance learning. In Psychology and the Real World.
4. Vygotsky, L. S. (1978). Mind in Society (Zone of Proximal Development). Comfort/learning-zone model popularized by Senninger, T. (2000).
5. Harris, A., & Harris, B. (2008). Do Hard Things: A Teenage Rebellion Against Low Expectations. Multnomah.
6. Oaten, M., & Cheng, K. (2006). Longitudinal gains in self-regulation from regular physical exercise. British Journal of Health Psychology, 11(4), 717-733.