If you listen to the news, Bill Gates, Elon Musk, etc., you may be worried that artificial intelligence is going to take your job.
While it is certainly true that there are some promising advancements in AI technology in the works, I sincerely doubt that we will suddenly have mass unemployment due to AI taking jobs (I wish).
Silicon Valley entrepreneurs may be great at understanding emerging technology, but seem to often not fully understand economics and the work that mere mortal humans do daily.
Humans are versatile
Even if you design an AI that can drive a truck autonomously, will it also be able to inspect the cargo? Do a quality test? Walk into the store to remove cash from the vault? Measure the level of oil in a tank at a well? Stock the shelf at the grocery store? Check the dates on the beer at the store and pull the old ones? How about diagnosing and repairing an air conditioning system including crawling through an attic and manipulating wiring?
Sure, I bet we can design one-off robots or instruments to tackle each of these tasks. How long will that take? How quickly can they be scaled out? How much will they cost compared to humans and is the ROI compelling? I am certain some of these will be done – possibly all of them, eventually. But it won’t happen overnight.
The more likely scenario
AI will enhance the productivity of the existing workforce. As with many productivity enhancements since the beginning of time, fewer workers will be needed to accomplish a similar amount of particular work. This is good. This is why only a few of us live on farms today. And here we are in 2017 with low unemployment in spite of millennia of productivity enhancing tools.
Unless there is a breakthrough of true superhuman intelligence coupled with the computing power to effectively have millions of Einsteins working together (Star Trek here we come!), these changes will be incremental innovations.
Individual forward-thinking companies in various industries will find ways to apply AI and robotics to improve some of their processes and decisions. This will give them an edge over their competition, which will then, in turn, employ similar technology. This will lead to increased profit for early adopters and eventual margin compression and falling prices as competitors get with the program – increasing everyone else’s purchasing power.
Take trucking for example. According to this report, fuel is 39% and the driver is 26% of the cost of running a truck. If Tesla is successful with their promised electric autonomous semi truck, then the fueling cost of trucking will fall dramatically and driver expense will be reduced (though not eliminated – they do a lot of work other than just drive). Tesla will only be able to build these trucks so fast. Trucking companies who buy them first will save money and increase earnings. Other trucking companies will follow suit when they see those profits. Eventually, those cost savings will likely lead to reduced shipping rates for the types freight that this technology is a good fit. Cost does not determine price, but in a highly competitive market like trucking, margin compression is to be expected as other firms implement the new cost-reducing technology.
How to invest
There will probably be a few hot companies that do really well with AI. It is hard to pick who those will be. However, it’s reasonable to expect that if AI delivers on its promises, many companies will earn more as a result of employing productivity enhancing technology, such as that trucking example. The companies that pay those trucking companies will also save money.
Maybe it will be data analysis to support better decisions, or certain physical robots, or cost reduction from needing fewer workers to accomplish the same task. This means that while some hot tech companies like Google, Apple, Tesla, IBM, etc. may do well as a result of creating this technology, many other companies across the world that aren’t tech companies will reap benefits.
There will be widespread increased profit margins and earnings for the early adopter firms that properly utilize AI.
I believe this is a strong case for owning something like VTI – Vanguard’s Total US Stock Market Index. I doubt the smartest and most well-informed expert on AI on the planet would be able to predict how this will shake out at a company specific / stock picking level. Some companies will flourish as a result and some may fail as their competitors eat their market share due to better product quality and/or lower prices. If you do try to pick individual stocks, try to find out if management at those companies is making proper use of productivity enhancing tools. This should be done regardless of AI.
Maybe some foreign competitors will beat US firms to the punch. It’s probably a good idea to own some VXUS just in case – even if they don’t outperform US companies, they will also benefit from utilizing AI. If a company like Hyundai uses AI significantly better and has some kind of breakthrough that lets them sell cars at much higher quality and much lower prices than US automakers, that VXUS would sure look nice.
Remember how you were supposed to be 3D printing everything at home now? 3D printing (so far) has matured into a great technology for manufacturing companies and a hobby at home. During the beginning of an emerging technology, people come up with fantastical dreams of how it will end up. I think AI is similar. We’ve been using computers to help analyze data and make choices for decades. That’s probably going to keep improving for a long time, leading to incremental and significant improvements in productivity and increases in our quality of life as a result.
There won’t be a robot dystopia or utopia any time soon.
Put most of your money in broadly diversified funds. If you want to try to pick and choose a few companies that you think will do really well from this revolution, buy some stock but don’t get too greedy.
But… what if it does happen?
You would still want to own VTI. If someone like Amazon does take over the world by using robots and AI to produce everything, throwing every human out of work, guess what? If you own VTI, you’ll own Amazon through it and be entitled to your share of the output from those machines. In fact, VTI would automatically consist almost entirely of Amazon, or whatever Mega-robot-corp-overlord emerges if that were to happen; unless somehow it was a private company, which I think unlikely.