The Erosion of Skill: Technology, Automation, and the Changing Nature of Human Expertise in the AI landscape

[This article was written completely by human hands and human mind. No AI was used (other than spell-checker, which I guess is a form of AI)]. All images are from Shutterstock photos taken/created by real humans, and used with permission].

You know you’ve been there. You’re driving through a town you’ve lived for 10 years, but you’re going somewhere new – perhaps to visit a friend at their home for the first time, or something like that. You put the address in your GPS, to make sure you find it. You follow familiar paths for a while, but, eventually, you end up a part of town you’ve never been before. Thankfully, your trusty GPS is keeping you on track with step-by-step directions. Nothing to worry about.

But imagine your GPS just stops working for some reason. Your phone dies or falls off the dash, and for some reason you’re no longer getting those directions. What would you do?

And be honest with yourself? Would you be able to, say, pull up a road atlas and find your way? Would you even be able to locate the destination address, without GPS pinpointing it?

50 years ago, most experienced drivers could have, relatively easily, gotten anywhere they needed by using a combination of road atlases and asking for directions. Now, though, we’ve grown so accustomed to our little computers telling us exactly where to go that we’ve lost our sense of direction, to an extent, and lost our ability to visualize, on a mental map, where we are and what direction we’re going – even in familiar places.

And this is just one example of how a technological advancement has eroded an essential skill we all once had. Now, I’m not saying that 50 years ago EVERY person could have found their way through unfamiliar roads using maps and intuition, but, compared to now, the average person was far better at this skill 50 years ago than we are now.

Let’s explore some other examples. I’m a musician and producer too, so I’ll use a music example. In the 1970s (and before), in order to record a piece of music, you needed musicians who were so practiced at their craft that they could play through a song with minimal mistakes. Music was recorded to tape, and editing tape was a real pain in the derriere. If recording engineers needed to take something out of tape, they had to literally find the spot on that tape they wanted to delete, and they had to either scrub it or cut it and tape it back together. The precision required to do this was tedious and difficult. Even that editing required a high amount of skill. So, musicians who could not produce a good performance consistently would not have been awarded much studio time with producers/engineers, because poor performers would waste too much time.

So, to be successful musicians, musicians had to practice a LOT. So did the engineers/producers. If your job is to record a live performance and get the best recording you can, you had to be on your game too to make sure you don’t ruin a perfectly good performance because you made some technical error (mic placed poorly, some intern interrupted the session, etc.).

In the modern world, with digital recordings and the availability of the home studio, musicians no longer need this same level of skill. I’m speaking from personal experience. I’ve recorded many, many things that, if asked, I could not play again in the same way, because I never had to practice enough to really learn the part. Additionally, I never had to get good enough at a part to play it perfectly, because I can just digitally edit notes/parts when I need. This meant I spent years as a recording musician who was just “okay” at several instruments. Don’t get me wrong, I’m fortunate to live in a time when someone like me can record music without paying hefty studio fees or winning a record deal, but I also know that I would’ve made better recordings if I had taken the time to develop better skill (something I’ve since remedied, by the way, in the last 5 years or so, by actually practicing and building those skills).

Modern musicians don’t even need to know music theory. They can create a whole song using virtual (computerized) instruments, and there are many (virtually unlimited) virtual instruments where a songwriter/player can literally hold down a single key on their keyboard and the computer/instrument will generate a melodic or rhythmic line form that single key. Recently, AI-generated music has gotten really good too. There are AI-generated artists on Spotify right now that many people have no idea are not real humans at all.

Part of the reason many modern listeners can’t tell the difference in AI-generated music and other human-generated modern music, in fact, has to do with the fact that a lot of modern music (even music made by successful artists) is not made by people who would have, in the past, been considered “professional musicians.” I don’t mean that these people are not talented, but I would say that their talent is not in classical musicianship or compositional skill. It is, instead, in making consumer-ready products, having charisma, connecting with audiences, and, perhaps, in vocal performance. I also don’t mean that all modern music lacks musicianship – it’s only a subset, but it’s a bigger subset now than it has ever been in the past.

Furthermore, ask any classically trained musician to listen to modern popular music and they will, most of them, tell you that the composition of modern pop music has deteriorated over time. Modern compositions are simple, almost always diatonic (all the notes fall into a single key, and never deviate from it), and follow a predictable, patterned structure – hook, verse, hook. Songs are also shorter (2-3 mins). The reasons music has gotten this way is complex and multi-factored – saying as much about listener and cultural trends as it does about musicianship. But, again, AI-generated music is better able to mimic this structure because there is such a clear, uniform structure, and, again, requires little actual knowledge of music theory or classical composition techniques. Computers are GREAT at following formulas, and formulaic music composition is a mainstay of modern pop music.

So why does all this matter. Well, again, I hope with this example, I’ve been able to lay a foundation to show how technology, over time, has contributed to a loss of skill-building. I’ve used music as the example above, because that’s an area/example I know well, and that I think will resonate with many people. However, hopefully you can all think of similar examples in your own lives, and the idea still permeates that as technology progressed to make life easier for people, the “easy way out” has also meant that people have not had to worked hard to get good at something.

And that is my great fear now that AI is evolving as quickly as it is. Why would we work hard to do something when a computer can do it for us faster, and with less effort? Everyone wants to spend more time “playing” than they do “working;” it’s only natural. But next I want to make some arguments about why we should continue to take the hard way.

1)    Efficiency is far from the most important human value. Computers outperform humans on efficiency, and if we prioritize efficiency too much, we run the risk of devaluing other concepts that, perhaps, should be more important to the human race. Some other values to consider might be, for example, humanity, morality, flexibility, open-mindedness, confrontation, kindness, creativity, and I could go on and on. If efficiency becomes the thing we prize the most, we will lose ourselves in our search to maximize it. And, consequently, we will lose to computers. If, instead, we put other values above efficiency, we can restore healthier order to ourselves and to the world around us.

2)    The “hard way” builds us into more rounded people. A lot has been said in the past 15 years or so about “grit” (just ask Angela Duckworth) and all the benefits it provides us to learn how to drive ourselves and our children toward a “growth mindset” (as opposed to a “fixed” one). “The hard way,” then, is a direct consequence of the growth mindset. Now, I’m not saying one should always choose the hard way, no matter what. That would just be stupid. But, I do think it’s important for people to take the time to “smell the roses,” that is, people need to hone in on the process of what happens as you go from point A to point B, and not get so focused on “just getting to point B.” That’s because there is often a lot of beauty along the “harder” pathway. And even if that harder pathway looks rough on the surface, it may give you opportunities the easy way never gave. That’s really what the growth mindset is about, then. It’s about taking the time to focus on the personal growth that happens as you journey through life. Because that’s really where the beauty lies, in the journey, not the destination.

3)    In this case, the “easy” way might just destroy the planet. There have been many reports now about the environmental impact of AI data processing centers. In short data centers are estimated to rank 5^th worldwide in electricity consumption, right between the countries of Japan and Russia. In other words, worldwide data centers use so much electricity that the electricity use of only 4 entire countries ranks higher (and those countries are China, US, India, and Russia), and the U.S. alone is responsible for about 9% of all global data center electricity consumption (second highest is UK, at 5.1%, followed by the EU, at 4.8%). So, data centers within US soil are absolutely guzzling energy, and this has a profound environmental impact (see linked article above for more info on that).

Data centers also use A LOT of water, up to an estimated 5 million gallons per day, which is roughly equivalent to the water use of up to 50,000 people in an average U.S. town. They’re using freshwater too, most of it to cool their systems. Some emerging data are also coming out which suggests that data centers are being built increasingly in places that are already somewhat water deficient. Consequently, they are placing a major strain on limited water sources. In 2022, an Oregon-based Google data center alone used 29% of the town’s entire water supply, and that is just one example.

So, if we keep using this technology “willy nilly,” we run the risk of killing our planet, eventually.

For what it’s worth, I used Google quite a bit to search some information on these data centers, and, in doing so, it occurred to me that now every time I use Google, I use AI, and there is no way to turn off those AI summaries that now appear at the top. So, I also conducted a little review on how a person could access a search engine that minimizes data center consumption (of course, you’ll always use some energy when you search, but if you can avoid generative AI, you will save a lot). Here are some of the search engines I recommend, based on my review. I’ll give the caveat that I’ve not personally tested all these, but am basing my info on publicly available policies and so forth.

Swisscow. This is one of the few search engines that does not default to AI summaries. It seems like you can click on an “AI summary” button if you want, but it’s at least not a default behavior. There are also some other privacy benefits (no tracking, totally anonymous) and it is family-friendly, excluding nefarious or “adult” sites (you know which ones I mean).

Qwant. This search engine looks a lot more like Google, so may be more familiar to Google users. You have to sign in to see AI summaries, so if you use it without signing in, you can avoid AI use.

Ecosia. Qwant and Ecosia recently partnered, so you are really using the same search engines when using both these sites. They are focused on European-centric search engines, trying to find independence from U.S. based companies (namely, Google and Microsoft). Ecosia also plants a tree for every 50 searches (and they plant them all over the world, not just in Germany, where the company is based). They also have a “beyond neutral” pledge, meaning they “produce more clean energy than it takes to power all searches and AI queries on Ecosia.” This is the one I ended up choosing to try out today. I’ve made it my default search engine (which is very easy to do). If I hate it, I’ll come back and edit this page to let you know why.

In sum, I hope this article inspires you to consider how you (and those around you) can move away from an overreliance on technology. Not only is it important for us to maintain essential skills, and it doesn’t just serve us as individuals, to be better, it also helps the planet. Sure, it might be easier and quicker to ask ChatGPT a question (I’m guilty of using it more than I should have), but ask yourself, is the “ease” worth the cost. For me, it is not. As of today, I’m pledging to avoid the use of AI when I can. If I do use it, it will be for specific, and very limited queries. Remember, there’s nothing AI can do for you that you can’t do for yourself (with more work) or ask another human with expertise to do for you. And when you do it yourself, you’ll have earned a skill you didn’t have before. And when you ask another human, you’ll have earned a relationship you didn’t have before. It’s a win-win (even for introverts like me).