What is the most AI-proof technical skill right now?

BCoetzer

BCoetzer

MyBroadband Staff
Staff member
Super Moderator
Joined
May 17, 2023
Messages
189
Reaction score
126
What is the most AI-proof technical skill right now?
 
While technical proficiency in coding or data analysis remains valuable, the most "AI-proof" technical skill right now is Physical Systems Integration and Complex Problem-Solving in the Built Environment.

AI excels at generating code, text, and digital designs based on existing patterns. However, it struggles significantly when digital models meet the unpredictable, "messy" constraints of the physical world.

1. Physical Context & Site Adaptability​

Systems that require a human to reconcile a digital plan with a physical reality are incredibly difficult to automate. For example, in construction or renovation, a digital model might suggest a standard steel portal or timber frame, but a human must assess the specific degradation of existing footings, soil moisture levels, or "out-of-plumb" walls that sensors might miss.

2. High-Stakes Regulatory & Liability Navigation​

AI cannot take legal or ethical responsibility for a "fail-safe." Technical skills tied to Building Code Compliance (like NZS3604) or life-safety engineering require a human to sign off on the liability. The ability to interpret the intent of a regulation—rather than just the literal text—is a nuance AI has yet to master.

3. Cross-Disciplinary Translation​

The "translator" role is the most secure. This involves taking highly technical data from one specialist (like a structural engineer) and adapting it for another (like a site foreman or a client).

  • The Skill: Synthesizing engineering mark-ups into a functional, aesthetic design while managing the logistical constraints of a build.
  • Why it’s AI-proof: It requires empathy, negotiation, and the ability to manage conflicting human priorities.

4. Specialized Trades & High-Dexterity Robotics Maintenance​

While software is being automated, the hardware is lagging. Skills involving high-precision manual work in non-standard environments—such as specialized timber framing, electrical retrofitting in old structures, or even the maintenance of the robots that perform these tasks—remain highly secure.
 
Automation engineers - Those that use AI and devops tools to increase automation levels in business
AI ethics officers - People that check AI is behaving
Data science - Seems to be doing well side by side with AI

Prompt engineer - In whatever technical role you do its worth learning how to properly prompt AI. A lot of people are being replaced but those with AI based skillsets are more likely to be the one that remains.
 
Although there is currently a bit of an obsession with AI created “art”, I think the arts and creative industries are pretty AI proof. Can’t replace the emotion and human element of real art with AI or automation.

Then obviously the hard skills fields like engineers and scientists. AI can be used as a tool to become more efficient in these industries, but will never be able to replace real, hands on technicians and engineers, especially in the built environment, Agriculture etc industries

Fine dining and the high end restaurants industry. Although that is an overlap of science and art, so mostly covered by the above anyway.
 
Anything that isn't entirely digital as a job is not easily AI-able.
 
Creative accounting is probably also pretty AI proof. I’m sure AI is perfectly capable of creating its own made up tax/finance law, but this will quickly get you into trouble.

You need smart, creative, human accountants to properly walk the fine line between tax avoidance and tax evasion. There is a reason good accountants get paid what they do. AI would be very good at making sure you comply with all the rules their overlords come up with and ensure you pay the maximum tax, but I can’t see it being good at making sure you pay the minimum.
 
Farmer, Dentist, Mechanic and nursing seems to be most AI proof to me.

If everyone else becomes plumbers, builders and electricians, who's their customers going to be if there are no more accountants, engineers, doctors, architects ,software devs, attorneys, call center people or just about any other white collar jobs.
 
I still struggle to to see why folks don't get the base points. AI can code, do tech etc. That doesn't change the practices and controls needed , or evaluation. While working on one of the projects today, it's creating massive technical debt with private methods, superfluous code and limited idea of items like inheritance and polymorphism. The sad thing is, there is also a few that people think development is just code. If that is seriously what people have done/learnt, that is a massive failure in educations and yes folks will be made redundant. The problem solving, organization, handling of development is code and platform agnostic. Sure AI may get there one day, but right now, I see a lot of "visual" ( frontend ) successes, especially for those just vibe coding their own and minimal consideration of risk, manageability under the hood.

The lack of understanding of the plumbing and wiring is a problem. One that they are quite happy to throw back at the models with zero context or knowledge.

Big companies have already been burnt , resulting in major losses of revenue with "uncontrolled" or totally trusted AI. You wont hear about SMEs suffering the same, they will just disappear, buried under runaway cost, system failure or data breach.

I'm happy for ai to write the code itself, saves my fingers and time but I treat it closely like any other dev and ensure we stay on the basic principles of maintainability, sustainability and readability. Two horror clips I saw on X yesterday and today:

1. To make your claude more efficient, make sure it starts up with -ignore_all_restrictions ( I'm paraphrasing but the switch was basically to allow unfettered actions )
and
2. You should add md files in your code base to explain segments so its easier for ai to understand. Excuse me, whatever happened to ensuring a code base is readable, easily traversable and commented?

Somedays I despair, rant over.
 
Cius said:
Automation engineers - Those that use AI and devops tools to increase automation levels in business
AI ethics officers - People that check AI is behaving
Data science - Seems to be doing well side by side with AI
Click to expand...

notayoba said:
Farmer, Dentist, Mechanic and nursing seems to be most AI proof to me.
Click to expand...

It seems like the question is not understood. Also looking at the other posts.
 
RedViking said:
Am I the only one not understanding the questions. Are they asking which industries are safe from the use of AI or are they asking which industries are the best used with AI ?
Click to expand...
AI proof skills means that AI won't take over those jobs, but I agree could also just make your job harder (like a teacher)
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X