The AI disruption matrix

We are living in interesting times, to say the least. While it is not new for technological progress to drive us forward towards new way of livings, the rate of change is unprecedented. It is only logical to wonder, what does this really mean for your, or any, job, let alone businesses and the economy.

My view on this, from the start was that most jobs are fine and AI is a tool ⚒️ but now that I have AI doing most tasks I was doing at my pre AI phase of my career, even I started wondering. However, I still think my original thesis stands, we just need to find a more timeless and grounded way of describing the argument. So here is the revision.

I suggest that they way you should assess the changes AI brings to the job market is around these three questions

😱 Is my job exposed to AI?

📈 Will there be more demand for my job?

💰 Will my job be worth less or more?

Is my job exposed to AI?

There are many ways to answer this question, but here is the framework I use

✅ Is the outcome of the job verifiable? think code that either runs or not versus a legal contract

👩‍🏫 Is the output an artifact or human delivery? think teaching compared to an Excel sheet

This creates the following four quadrants with the most exposed jobs being the ones producing artifacts (code, excel, accounting books) and are highly verifiable and the least exposed those that are not as easy to verify and need human delivery (teaching, medicine, coaching, therapy)

Will there be more demand for my job?

Interestingly high exposure does not translate to high disruption necessarily, something that is usually referred to as Javons paradox. This is the idea that as cost of production falls i.e. cost to produce code falls, demand increases, sometimes to the point where there is more demand than before. We are seeing this first hand with software engineering that instead of being replaced, even though AI is writing most of the code nowadays, its demand is only growing.

I have made that point in the past but we are now seeing that in action, as AI is doing more and more of the software engineering role, more people and business are enabled to write software but at the same time need help and accountability from an expert. That idea is not bounded to software only, there nearly enough doctors or lawyers as well as many other experts. It does not apply to all jobs though, there is only so much bookkeeping or architecture you can do.

Will my job be worth less or more?

This naturally leads us to the last important question which is around the value from any job given the AI taking over part of the role. This really depends on the part that AI is automating. Is it automating the high value part of the role or the other way around? We are seeing examples of both at the moment. In translation for example, AI seems to have taken a big part of the high value part, the actual translation, and left translator proof reading AI outputs. In software engineering on the other hand, it turned out that the high value part was the architecture, design and understanding of the system instead of the writing the code part which is also what the AI has automated.

The AI disruption matrix

Lets put all the above together in a matrix along the axis that matter i.e. will demand for job grow and will it be worth more.

The most important take away point for me though is that the AI disruption, if you noticed, is mostly happening within professions i.e. certain parts of a job like QA testing, drafting, bookeeping is going away but other parts or areas remain strong i.e. architecting software, accounting, strategising. So in most cases, there is plenty of room to move towards part of your job or expertise that is higher value and less exposed.

So in a nutshell, most jobs have and will continue to have parts that are high value and in demand and those are around the human parts i.e. judgment, taste, communication.