I had a conversation with a colleague recently about the future effects of Artificial Intelligence on the workforce. What we discussed ended up looking rather grim for mankind, but not in a sinister way. It was more a prediction on what opportunities would be available to us once the eventual AI renaissance begins AND how many regular jobs could deteriorate over time from its growth.
As we talked and talked, we both came to a few conclusions that we both agreed upon. Please note we are both not morons. Our agreed upon conclusions were:
Artificial Intelligence will become more widespread faster than we anticipate
It’s effects on jobs will be polarizing by:
eliminating jobs that can be automated away
forcing people to do jobs they were used to doing i.e. become entrepreneurs, go into the trades, etc.
Its’ effects on jobs will not reach its peak levels for another 10-20 years.
At some point, government intervention will occur. Similar to a Sarbanes Oxley, as something negative will spur this.
Those companies who do not adopt AI will struggle to compete.
At the end of the day, it will increase the gap between the haves and have nots. It will create opportunity for some and be the downfall of others.
While our Cohibas burned in the moonlight and our second round of McCallan neared its end, we started to dig in to what the logistics would be around how AI could actually eliminate jobs.
For most companies the effects would not be felt immediately. It would happen over time as the younger generations began utilizing AI tools. We speculated that most Boomers or older Gen Xers wouldn’t dare touch anything like Chat GPT. Why would they? They don’t even want to use Excel or Powerpoint. They can barely use Teams.
Now there are some small outliers where older leaders would heavily invest in AI from the start and reduce their workforce immediately. We’ve already seen it, however it is not wide spread, as AI is still in it’s infancy.
Most organization will adopt and delete over time.
We took a standard corporate finance organization as an example. Below you will a standard small to medium sized business finance org chart.
Our hypothesis was that over time, as more users adopt AI technology, that the need for backfilling positions will be less, especially as staff leave or age out.
We also made a few assumptions, knowing that this may be a rare case but we assumed that:
AI adoption would happen at the entry level to middle management levels. Those would be the core users, either directly or indirectly. (i.e. entry level staff utilizes AI or middle management reviews work of AI)
Headcount will never increase
AI will only increase capacity, never decrease it
AI will continuously improve
For the upcoming example, we also assumed that employees would not leave the organization and gradually ascend the org chart. (I know it is a bit unrealistic, but humor me)
Year 0
Given these assumptions, we drilled down into the basic accounting organization. Below is an org chart showing a very lean accounting org from the CFO down to a single staff accountant in year 0.
If we assume that only the staff accountant starts utilizing AI in present day, they are learning AI tools and adding that to their skillset. All of the higher level positions are not utilizing it at all. As the positions are vacated, and the staff accountant understands what AI is capable of, the lower level positions would get absorbed by AI.
The staff accountant is probably using something like AutoGPT just to scrape data for any manual journal that do not have a direct feed into the ledger and they are also optimizing the error report designs for control reliability, at a minimum.
Year 5
As time goes on and the staff accountant becomes more in tune with AI, moves up in positions with the organization (to Director) and AI tech improves, a scenario presents itself where the position they left becomes obsolete because they’ve increased their own capacity. Now this org drops from four to three.
The new Director (formerly the staff accountant) remains in a reviewer type role, but they review the work of the AI utilization they implemented. Thus the new “staff accountant” is AI technology.
We also assumed that all the positions have moved up (VP to CFO, Director to VP) because the CFO more than likely retired. Most execs average age is 55, which they were at year 0, and are now 60. Time to move on.
The new org chart is below.
Year 10
5 years later and we are at a similar scenario. The Director who started out as the lowly staff accountant, has now grown with AI technology and has ascended to become VP. All that is left of this accounting org is themselves and the CFO. The org is largely automated at this point, and there are very few staff left actually doing work. Most of them are just reviewers and overseers of AI automation producing most of the work. Do you see where this is going?
Once it gets to this point, smart individuals would review this org chart and prepare for anything catastrophic because there is really only one person who knows what is truly going on. That’s right! The VP knows “where all the bodies are buried” so to speak. The CFO relies heavily on them to review and maintain ALL the financial processes.
It is also possible that this scenario even, is unlikely, but it truly depends on the capacity of the VP and CFO to see if they can manage all of the AI processes. In this extreme case, let’s assume they think they can.
Year 15
Finally, the current CFO decides to leave and the VP is now the CFO . . . with ONLY AI staff. They manage processes, not people doing things. It gets to a point where their capacity is stretched and they need to hire someone to help manage this.
This is where it turns worse.
Since companies have become more reliant on AI, workers have been needed less in certain roles (accounting for this example). In the above scenario, the current CFO wants to hire someone to help, but who? There is nobody available internally since they never backfilled anyone and if we assume other organizations did something similar, their is nobody to source externally either.
Oh well, I guess they will have to hire someone without much experience. How about an intern or someone right out of college? I am sure there are plenty of people with accounting degrees who need a job. Wrong! Accounting degrees have dried up. Since nobody is hiring accountants anymore, nobody is going to college for it. People are only going after jobs that are available. Everyone is starting their own business or ripping heaters at a job site.
The current CFO is left not being able to hire any knowledgeable person for this and their doom is imminent.
Too Extreme?
I know the scenario I gave was very extreme (I was also a little drunk with my friend when we were discussing this) but I wanted to illustrate what potential problems there was with adopting AI technology.
So how can you prepare?
Learn AI Tools
What will differentiate you across competition is you ability to know and utilize AI tools effectively. Anyone entering the job market now will be required to know how to use some form of AI, depending on the level of security lockdown your company has. Larger companies are branding their own version of Microsoft’s copilot already. See Deloitte with DARTbot, that they are rolling out to their employees. QuickBooks even has Intuit Assist. It won’t be long until most companies have their own version of this or license a current version in the market, if they aren’t already.
Start Your Own Business
Based on the example of the accounting organization I laid out above, jobs will decrease in certain industries over time. One would think an increase in entrepreneurs and business owners would increase.
As someone who has a full time job and a side hustle, I use a few AI tools that increase my capacity and productivity. I am sure others have used more, but a few I am using or plan to use are (*Please note I have no affiliates with any of these)
ChatGPT - mainly for copy ideas, service offering discovery for my clients and research validation.
Midjourney - branding and copy.
Looka - logo and brand design.
AutoGPT - any automated tasks that just need done. I will be using this once I increase clients.
Bard - I have used this a little to test it’s capabilities. I use it similar to ChatGPT, but not consistently.
Learn a Trade
Many of you will look down on this, but I know plenty of people who make decent money running their own business as a plumber, electrician, HVAC installer, etc. If you can scale this type of business, you are sitting on a gold mine.
I have a few clients in this type of work and own one of these businesses. One of the main problems they have had is scale because of lack of reliable labor. The most successful can get a few good employees to help scale and then they can sustain their growth.
As an example, one of my clients started a plumbing and HVAC install company. he worked for a local shop and then went out on his own with another employee he worked with. He wasn’t able to scale until he could recruit a few more plumbers and techs form other companies. He also had six kids and three of them are of age to take jobs for him.
Over time, there will be more and more opportunity here, not only for people entering the trades but for people to own a trades business. At the moment, AI has not found a way to take reduce these jobs so it is a safe bet to at least look into getting knowledge around a trade or look to buy a service business if you can.
Although this has been an interesting rabbit hole to go down, this is not my only thoughts on how AI could change the workforce, but it is a possible scenario that I could see happening.
What the future holds of how the human race tries to tame AI is unknown and one can only speculate.