Artificial Intelligence (AI) is all the buzz lately, and rightly so. I suspect it’s a bigger deal than you think. AI isn’t just some fringe, hypothetical Star Trek tech for geeks and nerds to drool over.
Not any longer.
Prediction – AI technology will creep into every facet of modern culture and will completely transform vast industries such that we will not recognize them in 10 years, perhaps 5 years, 1 year in some industries. To the average person, this transition will feel like a sudden tectonic shift, with scarcely any time to react.
AI is moving so fast from concept, to reality, to practical applications, that entire industries will need less than 20% of their current workforce… and they’ll still be more productive than they are today.
Any profession that operates under a reasonably defined, logical rule set will be supplanted by AI bots.
Any job that involves the transfer of information or data is subject to AI replacement completely.
Any position that can use the phrase “How can I help you?” will be replaced by a bot.
I’m not saying this is bad, nor completely good. I’m just saying it’s going to happen, and it’s going to happen very, very quickly. In fact, it’s happening now.
In the near-future, people will say, “Remember just a few years ago when we used to talk to real people who worked where we shop? Those were the good old days.”
In 2007, the first iPhone launched. At the time, it was so cool, a huge leap forward in technology and design. Now, ~15 years later, that first iPhone would seem antiquated. Technology has come so far in such a short time, especially in communications.
Meanwhile, energy has become very inexpensive, digital storage capacity has become nearly free, and server computation capacity has grown exponentially. These trends are set to continue.
While AI applications are kind of cool now for demo purposes, we haven’t reached widespread adoption of AI tools in the general marketplace. But that’s all about to change. We are just on the verge of the tipping point. Even now, AI is beginning to integrate into commercial applications. In fact, you might be interacting with AI bots on occasion and not even realizing it.
As the AI gets smarter (it learns… and there’s an enormous amount of content online to learn from), like the embracing of the iPhone, we will see broad adoption of AI in the daily lives of regular people. It’s simply too good to not find its tentacles in every crevice of industry and leisure.
Jobs Where Bots Can Outperform a Human
Be careful what you encourage your kids to study in college. There might not be a job for that knowledge upon graduation. Here are some jobs the AI-enabled bots will outperform expert humans in short order:
- Professionals (Knowledge Workers) that operate within the confines of fixed language sets and information exchanges.
- Accountants – accounting is almost pure data and rule following. This industry is ripe for AI, especially for tax filings where AI will dominate humans in comprehending and resolving complexity, near instantly.
- Lawyers typically work within a standard set of language and ideas, often based on precedents in case law. Many legal documents are based on templates and standard language constructs. Almost the entire body of law is information based.
- Medical Doctors deal primarily in information discovery, draw conclusions from the given input data, and communicate selected outputs to their patients, based on test results and probabilities. In other words, doctors are in the information exchange business – information in, deduction and reason, information out. Further, doctors (especially specialists) operate within the confines of fairly standard and limited language. It certainly doesn’t feel confined nor simple when you listen to doctors speak. It sounds vast and complex. But, medical professionals generally stay within the realm of medical and healthcare words and phrases. There’s a limited number of medical things that can be wrong with a person. It’s not like a patient shows up to the doctor’s office to discuss jackhammers, concrete, and construction techniques. No, they talk about symptoms and things related to biology and health. So, if the universe of possible topics two people might discuss is focused on just one industry category or topic (like health), especially one that is canonical, like medical diagnosis, it’s a fairly restrictive, focused conversation (compared to a fully open-ended conversation about any random topic two people might discuss). Physician work is perfectly suited to AI applications.
- Teachers – I know there will be significant resistance to my stance here, but any subject that deals primarily in facts, data, information, and processes will be better taught by an AI bot. Think subjects like history, math, science, English, etc. The bots will simply outperform the very best human teachers we have… and it won’t even be close. They will know considerably more than their human counterparts. In fact, initially, you may not even know which teachers are human and which ones are bots. But, after some time, you will learn the difference… because the bots will be so much better.
- Customer Support – customer service jobs and call centers are also largely about answering questions within a narrow range of topics, words, and language. Nearly every job designed to help customers use or understand a product or service will be eliminated and replaced by an AI bot, who coincidentally, speaks at least 50 languages fluently.
- Retail – restaurant staff, anyone operating a cash register, etc.
- Imagine you work at Starbucks, at the drive through window. Today, you take customer orders through a crappy speaker-microphone pairing. I don’t know how much these jobs pay, but say it’s $15/hour. That’s ~$30,000/year for an 8-hour shift for that position. $60,000/year to cover two 8-hour shifts per day, excluding weekends. An AI bot, right now, can cover this position with ease. There’s a very limited number of total words, phrases and concepts used in these jobs. It’s mostly hearing and repeating the same things many, many times. Some AI licensing company says to Starbucks, “Hey, how would you like to go from paying $60,000/year to fill this position to $10,000/year with an AI bot?” The bot doesn’t get tired, does not take breaks, and never gets disgruntled. Consistent, hard work and output. No management. No payroll. No staff scheduling and management of shifts. Hyper efficient. You get the idea. “So, I can save $50,000/year per store and have better, more reliable customer service? Sign me up.” Suddenly, traditional entry-level jobs disappear across the country, across the world. Select AI bot licensing companies are set to make more money than any company in the history of the world.
There are real, considerable, broad-sweeping social implications here. There won’t be time to plan for drastic changes. It will unfold seemingly suddenly, as if the world changed overnight. At least it will feel that way to normal people. To the AI programmers who have been working on this tech for 30+ years, it will feel quite slow. But not for the rest of us. For us, AI will transform the landscape of work like nothing we have ever seen before.
My kids will tell their kids that they used to have to talk to a person to order food. There will be some future “old school” establishments that will thrive on the novelty of this concept… that you still order your food the old fashion way, by talking to a human, who physically writes down your order and occasionally screws it up, sometimes with a less-than-stellar demeanor. This will be nostalgic, like a 50’s diner is today.
New Stuff, Cool Stuff
The good news – after AI makes many things more efficient (simpler on the surface but more complex underneath… with the complexity hidden from us, as it should be), it will then be directed at creating new things, working on completely new ideas for products and services that we have yet to imagine to date. Some short time after some of these new areas launch, we will wonder how people lived before this magical technology was invented… much like we think of microwaves, air conditioning, internet, and cell phones today.
Non-AI-able Jobs (yet)
Some jobs are immune from AI intrusion (for now). These positions fall into some broad categories:
- Human Touch – Massage therapists. Chiropractors. Dentists. Nurses.
- Judgement – Management. Strategy. Leadership.
- Construction & jobs that require fundamental trade skills – plumbers, electricians, welders, ranchers (excluding farmers), etc.
- Creativity & Movement – museums, dance performances, some live shows (although likely AI infused). Art, music composition & arrangement will, however, be swallowed by AI.
Of importance is the answer to the question: Where do we go from here?
This is especially important for young people, as they enter the workforce.
A common thread from both my experience and in conversations with others is that AI is still a tool. We prompt it with some input (perhaps a question or a request) and it spits out some form of an answer. Sometimes that answer is incredibly insightful. Sometimes it’s just suitable. Other times, it’s simply not true, off-the-mark, or nonsensical. This will change and improve. However, in the near-term, with several iterations of our input request to the AI service, we can usually devise an improved query such that the AI’s response is more useful.
Going forward then, a key skill will be the ability to properly query AI services such that you get a useful answer. For simple problems, this can be trivial. For complex problems, this is not as simple as it might at first seem.
Often complex problems must be broken into component parts. Not all AI algorithms are well-suited for every type of problem or request submitted to it. There will be an AI transition season (which we are entering now) where a distinctly useful skill will be the knowledge of which AI systems to string together to break a large problem into smaller, solvable problems, and then to assemble the wholistic answer back together for a comprehensive solution to the problem at hand.
To gain this skill, experiment with AI systems and become proficient in knowing how to ask the right questions such that you get the best answer, in the format you want. The only way to learn this is through trial & error and experimentation.
Especially for young people, (high school, college, and early career), I’d recommend spending significant time on this. Some people will get AI to perform productivity miracles for them. These people will have productivity on the order of 10x – 100x of those who do not have this skill. Technology leverages and scales like that, especially AI.
The internet has significantly changed how we work (more efficiently), but that’s nothing compared to what’s coming. The rapid advancement of AI technology and its potential to transform industries is unparalleled. The social implications of these changes will occur suddenly with little time for acclamation, both for the leap in productivity and the creation of new products and services we have yet to imagine.
Young people, especially, should focus on mastering AI input query skills to enhance their productivity in the future.
Personally, I’m looking forward to it, even if somewhat trepidatiously (I recognize that bad actors also gain more technology leverage through AI).
By the way, AI suggested the title for this blog post, which was better than the title I had written.
Follow Past Midway if you would like an email notification of new posts.
I was in a time crunch a couple of weeks ago when I received a request from an employee urgently needing a recommendation letter for a business school he is planning to attend. I knew what I wanted to say, but lacked the time needed to polish the language with an appropriate tone. I had just set up an account with ChatGPT so I decided to give it a try. I wrote in colloquial terms what I needed. It took about two minutes to think through what I wanted to say and type the request in about five sentences. The result was a six paragraph letter that said exactly what I wanted and in the right tone. Copy, paste. One of the paragraphs was a bit too formal for my taste so I summarized the idea into one of the other paragraphs and deleted that paragraph. I was able to produce a signed letter in about 10 minutes, start to finish. On my own, this would have taken me 30 minutes of my own time (slow writer), spread out over two days between meetings and other tasks.
The employee later thanked me for the letter and said it secured him an additional $14,000 in scholarship funding.
The ideas were my own, set in the language and tone I eventually could have produced, but the detail work was performed by AI.
This was a win/win.
Well written as always. The teacher, lawyer, doctor points were interesting to me. Certainly had not considered lawyer or doctor. I’m going to have to give that some thought.
Teaching, I agree, but I think it depends on the age, and it relates heavily to my thoughts that education is (has been) due for a great overhaul / reimagining. I don’t see A.I. being able to manage a classroom. And we have already seen that remote learning has many many downfalls. I am wondering if we may shift the idea of teacher, to be more like babysitter. Which I would be fine with, if the requirements for that role actually focused on child psychology, classroom management, outdoor play, helping students with tech in the classroom, etc. We are already at the point where we need to realize that homework is obsolete and testing needs to be done in person because students are using AI (like all tech) much faster and more efficiently than their teachers.
It’s going to be an interesting ride.
As always, you write thought provoking articles.
I just have a hunch, AI is going create more jobs, and going to require more skills from existing professions.
1) When I was in high school (early 60’s), India banned computers b/c it would replace labor.
2) When IBM introduced IBM 360 in 70’s (I was in US), the prediction was that the accounting profession will be wiped out b/c companies will have financials at 12:01 AM every month, every year.
3) When Edison showed electricity transmission, NY Times front page said this tech will kill humans.
I am sure there are other predictions of tech’s negative impact. In each case, need for higher/faster production has outweighed tech’s productivity gains.
In case of AI, I think it major weakness is input quality. We all know GIGO (Garbage In, Garbage Out). AI is already able to weed out junk info, but going forward, I think humans will intentionally add more junk to the inputs.