One of the many fears people have about artificial intelligence (AI) is that it will take their jobs. From screenwriters to longshoremen. there have been several protests and labor actions against the use of automation and AI. But this fear may be unfounded in the electric industry due to economic laws of supply and demand. Simply stated, there is a lot of work to do in the electricity industry, and fewer people to do it. If nothing else changes, the cost of labor may rise so much we cannot do the needed work. Can AI save the day?
The amount of work to do in the electric industry seems overwhelming. Electrical demand is increasing quickly due to rising standards of living in the developing world, the shift to electric vehicles in the developed world, and growing demands for data centers. The nature of how we make and deliver electricity is also changing. The underlying fuel for power plants is shifting to include more carbon free or neutral sources. The electrical grid itself is moving away from relatively few large central power stations running all the time to an array of smaller generators with widely varying outputs. An EPRI report finds that worker shortages are likely for a variety of scenarios. The report states “those trades where worker shortages are most likely are boilermakers, carpenters, insulators, ironworkers, and millwrights.” These fields are critical to electrification, carbon reduction, and meeting increased electricity demand.
At the same time, there are simply fewer people to do the work. Birth rates in most of the world are falling. The reasons for this are debated, but there is no question it is happening. For example, a fertility rate of 2.3 babies per couple is estimated to keep a population stable. South Korea now has a fertility rate of 0.72. The G20 has an average fertility rate of 1.6 (https://www.worldeconomics.com/Regions/G20/). Many other countries are showing low fertility rates as well. Where once some were concerned with overpopulation consuming the world’s resources, now there is legitimate concern that we will not have enough people to produce the resources we need. And the people are aging and may soon leave the labor market, being unavailable to train their replacements.
All else being equal, the result will be an increase in the cost of labor and concern that there will be a shortage of trained workers to do the work that needs to be done. But, if we can make each worker more productive, then the higher cost of labor could be sustainable. It’s not about the money. It’s about productivity.
A simple analogy – we need to dig a ditch 1,000 feet long, but there is only one person to do it. Is it better to tell them to start digging, or figure out how to maximize this person’s productivity? AI can help with that.
For the workers we have, it may be better to train them to invent machines vs. doing work. Procedure writers could learn how to use AI to become more productive in developing work instructions. Power plant mechanical workers could become technicians for automated controls. Although AI cannot turn a wrench (yet), AI could improve planning and scheduling to maximize “wrench time” (the time a worker spends working vs. the time spent training, planning and preparing) in the field. AI could also help focus scarce workers where they are needed most.
For example, nuclear plant workers must spend a lot of time inspecting components and systems. Most of the time the equipment is fine. AI could identify the components that are fine, which allows inspectors to focus their scarce time on items that require human critical thinking and judgment (see for example, “AI-Assisted Analysis of Ultrasonic Inspections”).
AI could help teach workers new skills. An exciting use case for AI is to develop individualized curriculum for learning new skills. No matter the population, we will need to train new workers or teach workers new skills. Typically, when you start a new course of study, you already know some things about the topic and may pick up some concepts faster than others. An AI program can test your knowledge up front, monitor your process, cover the things you are good at more quickly, and focus on the areas requiring more time. This individualized learning is achievable now.
The bottom line is that we need to welcome AI (with the caveat that it is equitable and accessible) to facilitate the electricity transformation that must happen and the grid to support it. In fact, without AI, the world could face dramatically rising labor costs and declining productivity. Instead of fearing AI, we need to embrace it and its promise to automate and improve workers’ lives. For more information, see https://www.epri.com/thought-leadership/artificial-intelligence.