TL;DR: A crutch isn’t always bad — sometimes you need one to heal. But when you use a crutch indefinitely for a leg that could walk, you weaken the leg. This article explains when and how AI crosses from amplifying your capabilities to replacing them, and what the long-term cost looks like.
The Short Version
There is a version of AI use that makes you genuinely more capable. You use it for the right tasks, at the right times, and your own skills compound underneath it. You get better at prompting, better at evaluating outputs, better at directing AI toward high-leverage work.
And there is a version that hollows you out slowly. Your surface output goes up. Your underlying capability erodes. One day, you sit in a meeting without your laptop and realize you can’t hold your own in the conversation the way you used to.
The difference between these two trajectories isn’t how much you use AI. It’s whether AI is serving your growth or replacing it.
The Tool vs. Crutch Distinction
A tool extends your capability. A hammer lets you hit harder than your fist. A calculator lets you compute faster than your brain. The key property of a tool: you remain the agent. The tool serves your intent. You direct it. When it’s not there, you are still capable — just slower or less powerful.
A crutch substitutes for a capability you have or could develop. It takes over a function. And when it does, the underlying capability atrophies from disuse.
📊 Data Point: Sports medicine research on injury rehabilitation shows that crutch use beyond medical necessity leads to measurable atrophy in the supported limb within 2–3 weeks. The same principle applies to any system that bears load on your behalf.
The question is not “am I using AI?” The question is “is AI bearing cognitive load that my brain should be carrying — and am I weaker for it?”
Three Categories of AI Use
Category 1: Amplification (tool)
AI handles tasks that are genuinely outside your core value-add: formatting, research synthesis, boilerplate generation, translation, summarization of external content.
You remain fully capable of the core work. AI just removes friction. You would grow exactly the same without AI in these areas — it just takes longer.
Category 2: Scaffolding (tool, with awareness required)
AI helps with tasks that are within your capabilities but where it adds genuine leverage: exploring options, challenging your thinking, catching errors, providing structure.
This is valuable — but it requires awareness. If the scaffolding becomes permanent rather than temporary, if you stop being able to build the structure yourself, it has crossed into crutch territory.
Category 3: Substitution (crutch)
AI performs work that you should be developing the ability to do yourself — or that you used to do well and have stopped doing.
💡 Key Insight: The substitution category is particularly dangerous for skills you’re supposed to be developing, not maintaining. If you’re early in your career as a writer or developer or strategist, and AI is writing your first drafts or architecting your systems, you may never develop the baseline skills that make you genuinely capable in your field.
How to Find Your Crutch Lines
The crutch line is personal. It depends on what your work is, what you’re trying to develop, and what you need to own.
A useful exercise: answer these questions honestly.
What would I be unable to do if AI disappeared tomorrow? List these specifically. Some items are fine — you shouldn’t need to remember every regex syntax. Others are alarming — you shouldn’t need AI to think strategically about your own company.
Which of my current outputs could I not explain or defend in depth? If you’ve shipped something — a strategy doc, a piece of code, a piece of writing — and you couldn’t explain every decision in it from first principles, AI may have done more of the thinking than you realize.
What skills am I not developing because AI is doing that work? This is the hardest question. Not what you’re outsourcing today, but what you’re failing to build for tomorrow.
The Compounding Cost of the Crutch
📊 Data Point: Skill development research shows that competence in complex domains follows a compounding curve — early investments in deliberate practice yield accelerating returns over years. Interrupting practice in the early phases is significantly more costly than interrupting it at higher skill levels.
The crutch cost isn’t linear. It’s compounding in reverse.
If you’re a junior developer who never learns to debug without AI, you don’t just lack that specific skill — you miss the pattern recognition, the mental models, the intuitive understanding of systems that comes from wrestling with problems directly. Those are the foundations on which senior-level judgment is built.
If you’re a founder who outsources all strategic framing to AI, you don’t just produce lower-quality strategy today — you fail to build the strategic intelligence that should be compounding across every decision you make.
💡 Key Insight: The crutch doesn’t just prevent recovery. It prevents the growth that would have happened without the injury.
When It’s Okay to Use a Crutch
Let’s be honest: sometimes a crutch is appropriate.
If you’re under extreme time pressure on work that isn’t in your growth zone, AI substitution is a reasonable choice. If you’re executing rather than developing, outsourcing execution is smart. If you’re doing work that isn’t core to your expertise or trajectory, efficiency matters more than growth.
The problem isn’t using AI as a crutch sometimes. The problem is using it as one always, in all domains, without awareness of the cost.
What This Means For You
Do one thing this week: identify one thing AI is doing for you that you should be doing — or developing the ability to do — yourself. Not everything. Just one thing.
Then do that one thing, without AI, once a day for a week. Notice what happens to your capability, your confidence, and your relationship to the work. That’s your data.
Key Takeaways
- A tool extends capability; a crutch replaces it — the distinction depends on whether your underlying skills are growing or atrophying
- Three categories of AI use: amplification (safe), scaffolding (safe with awareness), substitution (risky)
- The cost of the crutch is compounding — it’s highest when it prevents early-career skill development
- Appropriate crutch use exists; the problem is unconscious, habitual substitution in your growth zones
Frequently Asked Questions
Q: How do I know if I’m in the substitution category? A: The test is capability, not frequency. If you stopped using AI for a specific task, would you be able to do it? If you could once but can no longer? If you’ve never been able to because AI was available before you learned? These three situations each warrant different responses — but all three are worth examining.
Q: Isn’t it fine to outsource things I’m not good at? A: Yes, for tasks that aren’t in your growth trajectory. No, for tasks where your growth is the point. The distinction depends entirely on your role, your career stage, and your goals — not on a universal rule about AI use.
Q: What if AI is just better at everything than I am? A: At execution of many tasks, yes. At judgment, contextual wisdom, novel strategy, and knowing what actually matters in your specific situation — no. Those are the capabilities worth defending. They’re also the ones that make you able to use AI effectively rather than just frequently.
Not medical advice. Community-driven initiative. Related: The Psychology of AI Dependency | Human Skills AI Cannot Replace | Reclaiming Creativity from AI