Are there specific industries or job functions that are more resistant to AI automation in the medium term?

You're looking for a safe harbor, aren't you? You're scanning the horizon, trying to find the one industry, the one job function, where you can ride out this AI storm without having to fundamentally change how you work. You're hoping there's a corner of the economy where the old rules still apply, where your current skills will remain valuable for at least the next three years, untouched by algorithms and automation. It's a natural question, a very human one, born from a deep-seated desire for stability in a landscape that feels anything but stable.

Here's the problem: that mental model of "resistance" is fundamentally flawed. It's like asking which parts of the coastline are "resistant" to a rising tide. The tide is coming in, period, full stop. What you're experiencing isn't just about specific tasks getting automated; it's a redefinition of value, a shift in what it means to be productive, and a fundamental change in how work gets done across every sector.

The fact of the matter is, there are no truly "resistant" industries in the medium term. There are only industries and job functions that are slower to adopt or where the impact looks different. But even those are temporary states. The hidden mechanism at play isn't just AI's capability; it's competitive pressure. If one company in a "resistant" industry figures out how to leverage AI to cut costs, increase output, or innovate faster, every other company in that industry must follow suit or die. This isn't a choice; it's an economic imperative. So, while you might find a niche where AI isn't directly automating your specific keystrokes today, the strategic decisions being made at the top, driven by AI's potential, will reshape your role, your team, and your entire organization whether you like it or not.

So, if you're holding onto the idea that your industry—be it plumbing, nursing, or specialized artisanal craft—is somehow immune, you're clinging to a false comfort. You're waiting for someone else to draw the line in the sand, to tell you where the safe zones are. You're expecting your current skillset to be a shield. But that's not how this works. The risk isn't just direct automation; it's obsolescence by indirect means. It's the new tools that make a team of five as productive as a team of fifteen. It's the AI assistant that handles 80% of the routine inquiries, leaving the human only the truly complex, empathetic, or novel problems – problems your current training might not have prepared you for.

What that means is, you can't afford to wait for a "resistant" industry. You have to become resistant to obsolescence yourself. You have to get on the front side of this wave, not wait for it to crash over you.

Here's your practical ladder, your next steps, starting now:

1. Identify Your "AI-Adjacent" Skills: Stop thinking about what AI can't do. Start thinking about what AI needs you to do. AI needs direction, refinement, ethical oversight, creative problem-solving, and the ability to translate complex human needs into machine-readable instructions. These are "AI-adjacent" skills. If you're a nurse, it's not about AI replacing your care; it's about using AI for diagnostics, scheduling, or patient monitoring so you can focus on the human connection. If you're a plumber, it's about using AI to optimize routes, diagnose complex system failures, or manage inventory.
2. Become a Director, Not Just a Doer: Your value is shifting from doing the work to directing the work. Learn to prompt, to refine, to critique AI outputs. Think of AI as a junior assistant that can do a massive amount of grunt work, but needs your expertise to guide it, correct it, and ensure it's actually solving the right problem. This isn't about being a programmer; it's about being a commander of AI tools.
3. Build Proof, Not Just Resumes: Stop updating your resume with buzzwords. Start building proof. Pick a problem in your current role or industry that AI could help solve. Then, use readily available AI tools (ChatGPT, Midjourney, CoPilot, whatever) to build a prototype solution. Document the process. Show the impact. This isn't about getting permission; it's about demonstrating initiative and capability. Proof that you built it. Proof that it works. Proof that it made an impact.
4. Teach Someone Else: The fastest way to solidify your own understanding is to teach it. Find a colleague, a friend, or even just document your process for yourself. Explaining how you used AI to solve a problem forces you to articulate your understanding and identify gaps.

What are you waiting for? Like literally, what are you waiting for? The idea of a "resistant" industry is a fantasy. Your agency, your ability to adapt and direct these tools, is your only true resistance. Start building that today.