When we automate friction, we accidentally automate away the learning.
AI may not steal your job. It steals the struggle that was quietly making you better at it, and hands you the result without you in it.
Every human task carries developmental friction. It is the productive resistance that builds skill, judgment, and expertise. Studies show expertise requires iterative failure cycles, not just output. When you delegate a task to an AI agent, you don't just remove the effort. You remove the neurological event where competence is formed. This is called the Human-AI Paradox or Human Agency Paradox.
So does it mean you should be wary of using AI, as it may leave you dumb? Well, no. I am not saying that as I don't want to be killed by AI fanatics :-). I am just an AI Advocate.
Nevertheless, we must appreciate that while not all friction is worth keeping, developmental friction certainly is. Debugging a production failure makes us better in understanding the system architecture and nuanced integration points. Drafting an outline of an article, book, or of a blueprint helps you research and learn about the new space. Writing a design doc surfaces contradictions that you may not have even noticed that they exist. These are all examples of developmental friction that, when we go through, makes us ready to prepare for the future.
So, a useful thumb-rule that you may follow: automate tasks that repeat without teaching, but preserve those tasks where the overcoming difficulty itself is the curriculum for graduating to the next level. Every day when I speak with executive leaders or observe when I am not speaking with them, I find umpteen examples where organizations are autonomizing process steps that used to create domain experts and business leaders. IT doesn't come without a cost that they can't see right now.
While efficiency is a great goal, what you might be removing efficiently could be the better version of yourself.

Comments
Post a Comment