How To Protect Your Job From AI Automation

AI’s coming for white-collar jobs isn’t new thing. Even in 2015, Martin Ford’s The Rise of the Robots foretold how AI will automate white-collar work — read chapter four if you want to put yourself some sort of midlife career crisis.

The problem isn’t AI’s coming for knowledge work because this is how society progresses like how the car replaced horse and cart, but this took a decade. The problem is that it’s happening sooner than what we would’ve thought. It doesn’t feel gradual — like something we can upskill and prep for. Rather, knowledge automation is happening now and requires our immediate attention.

The release of ChatGPT in November 2022 and the ever-burgeoning list of large language models make it seem like that tech companies are bent on automating knowledge workers despite trying to say that AI is to augment knowledge work. For instance, there’s a surge of tech layoffs for AI automation that’s happening now! This is nothing like the gradual transition from horse to car in the early 20th Century.

So, this has got me worrying — what can I do about it? I’ve got a family to feed and I’m useless with a hammer. If AI is going to automate my white-collar, developer job, then I’m screwed. In a moment of anxiety, I went online to find some ‘advice’. I felt more despondent after reading what was online.

All the advice just felt so conventional:

• Be friends with AI. Work with it. Not against.
• Strengthen your emotional and decision-making skills because they can’t be automated.
• Consider specialising in a niche that is difficult to automate.
• Be a supervisor of AI outputs. Learn how to wrangle AI.

All this information is well-meaning except if everyone is doing it then AI may as well automate it — you have to know that AI works by emulating patterns. So, if a certain pattern exists, AI will be able to automate it.

So, what can you or I do against AI automation? The reality is this: If you want to avoid AI from taking your job, you need to work in unconventional, chaotic and uncodifiable ways.

If you need an analogy for what I mean, think about the Story of David and Goliath. (1 Samuel 17) How did the smaller and lesser-equipped David defeat Goliath? He didn’t follow the ancient rules of military engagement. Instead of armour and a sword, he went into battle with great faith and a slingshot.

How could Goliath go against being shot in the forehead with a rock? He couldn’t. Goliath was expecting a melee encounter but he got shot by a miniature missile. This is asymmetrical warfare 101 or perhaps we could say guerrilla tactics. In the same way, we need to use unconventional tactics to protect our work from AI automation.

Don’t be a cyborg. Be a guerrilla!

Robocop is about a cyber-enhanced cop who dishes out justice against crime in Detroit. The cop, Alex Murphy, is murdered by a gang but is revived by the Omni Consumer Products megacorporation as a cyborg.

In the movie, the public is made to believe that humans are in control of robots when in fact the robot is controlling the humans on behalf of the evil megacorporation. It’s a 1980s classic.

Despite the 40-year-old premise, it somewhat echoes how technology drives us today. For instance, a study by Michael Gerlich at SBS Swiss Business School found that increasing reliance on AI likely diminishes skills in critical thinking. I can vouch for this. I dislike writing difficult regex so I get ChatGPT to write it for me. But, if you ask me to write a complicated regex, I still won’t have an idea how to. (ie. I’ve transitioned from traditional coding through trial and error to vibe coding.)

On the other hand, it would be a bit silly to say that you should become a Luddite and avoid technology at all costs because really modern society’s productivity is driven by technological progress. The graph below just shows how quickly we’ve progressed technologically in the past 200 years.

A graph showing the accelerating rate of human technological progression. Credits: Max Roser — Our World In Data

So, what can you do? Like David, you should understand the rules that your opponent works by. In this case, AI (as large language models and machine learning) works by discovering patterns in data sets. If something can be codified, AI can find a pattern for it.

In other words, don’t make your job routine — make it random and chaotic, although this is much easier said than done.

After you’ve been in a job for a few years, you’ll realise that it becomes routine. For instance, I’ve been a developer for awhile. At the start, everything felt new, but after a few years of coding and discussing with stakeholders, the same sort of issues always pop up.

They may be worded differently but the issues are usually the same such as stakeholders’ always forgetting that one key requirement they always want despite how often you confirm requirements with them.

This is where thinking in terms of biology can help you in avoiding your work from being automated. Life is codified in DNA, but once in a while mutations are introduced into DNA because the biological photocopying machines fail due to age or harmful pathogens damaging such machines. This means once in a while we should introduce some mutations into our routine work patterns. For instance, some of my colleagues would say that I wreak havoc at work. Instead of striving to get current libraries perfect, I will go off and introduce new features just because someone had said something that they ‘would’ like something in passing. Basically, I produce more new code than fixing up existing bugs. It’s a risky practice.

I keep my colleagues on the edge, but features that customers don’t even know they want are already there. I imagine my service like going on Amazon. The website tells you what else to buy rather than buying what you think you want.

The unpredictability of your nature will keep AI from being able to automate your job although you probably won’t be hireable with such nature.

The late celebrity chef, Anthony Bourdain, says in his book Kitchen Confidential that he only hires chefs not for their culinary ingenuity but to follow his system of cooking. In other words, he wants to hire robots.

So really, your best bet to save yourself from automation is to first find a job where the boss rewards you for constant innovation rather than being good at the everyday work.

Predict Where the Bottlenecks Will Be and Avoid or Exploit Them!

The problem will all these AI gurus telling us how to work with AI is that it leads us to herd mentality. When everyone listens to the same conventional advice and does the same conventional thing, then the system breaks down.

Martin Ford’s The Rise of Robots explains it like this: When AI takes over knowledge work, then conventional wisdom tells us that we need to uptake more upskilling and more education. However, there’s a problem here. AI reduces the number of ‘human’ positions available. The positions that remain pay more and require more specialised knowledge but there’s fewer of them.

Then, the law of preferential attachment (or winner-takes-all) tells us that only the best of the best or first movers will get the jobs and they’ll keep on moving up so leaving the rest of us behind.

Here’s my interpretation of Ford’s words. If you look at the workplace, a tech company may have a thousand developers, but only a tenth of them become middle managers, even only a tenth of them become senior middle managers, and only a tenth of them may only get into the C-suite.

Below is the organisational chart of a standard company which shows you just how hard it is to move up the corporate ladder.

A generic organisational structure showing the number of reports managers have as you read down the chain. Credits: Annemarie Gobin — Creately.com

As AI reduces the pool of roles for knowledge workers, your further education (if you’re lucky enough to keep your job) will only help you maintain the same role — albeit with a different title. It won’t necessarily help you move up the ladder as you’d think, and the chances of losing your job remain quite high. And, don’t think that those who already climbed the ladder will leave any time soon. Even they understand that preferential attachment works against them.

So, what can you do? Eliyahu Goldratt’s novel The Goal gives us practical advice on working with constraints and bottlenecks although in the context of manufacturing. Essentially, his book is about identifying bottlenecks, exploiting bottlenecks, supporting the bottlenecks and finding ways to expand capacity.

The point here is not to join the bottleneck (ie. follow the crowd). If everyone is taking on formal education and upskilling, that will create a bottleneck in the job market because essentially you’ll be competing against people with the same level of education as you.

If you’re from a standard middle-class family (eg. mortgage, kids and house) just think about how hard it was to land your first job. It’ll be the same except you’ll be an old person this time and ageism is a real factor in the workplace.

I’ve read that many places prefer to hire ‘new’ over ‘experienced’ just because ‘experienced’ usually means you’re resistant to change irregardless if that’s true to you or not. The common assumption is that old people don’t change or are too slow to learn.

Instead, I advocate looking for weaknesses in the system and positioning yourself as the go-to expert for that area.

For instance, my career evolved from being an admin staff at the company I work at to being a developer at the same company all without taking on formal education. I don’t have a degree in computer science — in fact, I’ve got a degree in pharmacy, which I left because I’m not fond of packing pills.

I got my role because at my interview said something like: “There’s a bottleneck of developers who understand the company really well and I can help you alleviate that issue because I under the business problems, and I code for a hobby — here see my work.” Admittedly, I never thought about being a developer because I don’t like to build stuff and my coding skills were limited to processing data faster, but I took on the opportunity because again there were bottlenecks in my desired career path — I wanted to be a data scientist yet I didn’t want to pursue a Master’s. Furthermore, I had a feeling that taking the same path as everyone else wouldn’t lead me anywhere because there were more people studying data science than actual data scientist jobs.

As mentioned, preferential attachment means that usually the best of the best or the first movers get preference. Everyone else is left behind waiting for an opening. I was neither the best of the best nor was I fast enough to be first in line.

However, by chance, I did become the best at understanding my company’s complicated business processes and being able to codify them.

If you’re a fan of the James Reece series (the book series The Terminal List is based on), the protagonist always talks about taking the high ground against opposing forces. This is more than just a military tactic. It’s about taking a position that enables you to exploit the battlefield as quickly as possible. So, if you want to compete for the diminishing number of jobs driven by AI automation, then start by taking up hobbies that put you on the unconventional path toward whatever career may be awaiting you because you just need to let preferential attachment work in your favour.

Conclusions

None of this to say you should be anti-AI. As a matter of fact, I used a few AI tools to help write this aruticle; for instance, ImageFX for image creation, NotebookLM for research, ChatGPT for writing evaluation and Grammarly for proof reading.

Rather, I think you should be focused on AI-automation proofing yourself, which involves a mindset of constant innovation and a desire to take upon yourself unconventional hobbies. Unfortunately, automation has shown to displace work, so it’s something you can’t ignore. For instance, Neil Dahlstrom’s Tractor Wars: John Deere, Henry Ford, International Harvester, and the Birth of Modern Agriculture tells how tractors and harvesting machines made farm hands and horses obsolete in agriculture.

I believe AI automation may do the same to us knowledge workers, so it’s paramount to figure out how you can be one of the few still with a job after the robots have taken over.

Read the full article here: https://medium.datadriveninvestor.com/how-to-protect-your-job-from-ai-automation-1a1be246ec5c