Fred Schindler
IMAGE LICENSED BY INGRAM PUBLISHING
Between high school, college, and graduate school I took lots of calculus courses. But how much did I use that knowledge outside of course work at those schools? Very little. In my professional career, working as an engineer, I’ve rarely employed any calculus. I can think of only a handful of times I’ve used it, and then only simple integrals and derivatives. That’s not to say calculus hasn’t extensively contributed to my work. But machines did it for me. Or more accurately, people skilled in the use of calculus wrote programs to perform calculations used in my work.
Electromagnetic (EM) simulators have become central to our work as RF and microwave engineers. The people who develop those tools use advanced mathematical techniques. As users, we should have a sense of what they do and how they work in order to make effective and efficient use of them. But the detailed mathematics is left to the developers.
Every time there is a new release of an EM tool, its developers employ new techniques and more advanced math to make the simulators more efficient and more accurate. They also automate more of the tasks needed to set up a simulation. I’m sure this trend will continue. I suspect the incorporation of artificial intelligence (AI) isn’t far off, if it isn’t already here. Every year the design process becomes more automated, which means the machines are doing more and more. A skilled engineer will still make better use of the tools than a clueless one, but might that eventually change too?
There was an interesting “Tech Talk” at the IEEE board meeting series in June. “Not Your Grandfather’s Tractor” was the talk given by Mark Moran, head of the John Deere Innovation Center. (You can view a recording at https://ta.ieee.org/meetings/ieee-meeting-series/june/tech-talk-not-your-grandfathers-tractor; IEEE log-in is required.) He spoke about the machines used in modern agriculture—tractors, combines, planters, etc.—and the automation that goes into them. Every new generation of farm equipment generates higher yields and uses fewer chemicals. The most recent generation of their equipment can operate without a driver in the cab. Farmers can run them from their mobile phones. The machines have taken over. Farmers without automated equipment can’t compete in commodity agriculture.
According to the U.S. Department of Agriculture, 1.3% of the U.S. population worked in agriculture on farms in 2021. I’ve seen estimates that in 1790 about 90% of the U.S. population worked in agriculture. According to the U.S. Census, about half the population worked agriculture in 1870, about one-third in 1900 and about a one-fifth in 1950. Clearly, farmers in the United States have become more productive over time. Technology has continually advanced. Does that mean that since 1790 89% of the U.S. population was idled, given that the machines have taken over?
Increased automation isn’t limited to agriculture or engineering. In the 1810s the Luddites fought against increased automation in the textile industry in England. They sabotaged the equipment that increased automation, that could produce textiles faster and more cheaply. Skilled textile workers were being replaced by machines that could be operated by less-skilled workers. The Luddites reacted to how these machines threatened their livelihoods, their way of life. I have sympathy for what they went through. But today textile production is highly automated, and no one is sabotaging equipment anymore. In fact, we have automation to thank for all the inexpensive clothes now we have available to us, for better or worse.
Over the past 200 years or more, we have seen a continuing trend of ever-increasing automation. The pattern has been that, even if there are disruptions in employment and social structures, other forms of work develop, and there is once again plenty of work to be done. No one wanted or needed a smartphone in 1810. There was no market for mayonnaise in a jar then either. At that time, most people only had one or two pairs of shoes. Dwellings were modest and lacked plumbing, appliances, and air conditioning. Lives were short and medical care was primitive. Automation has brought most of us a better standard of living.
So, if the machines have already taken over, have we nothing to fear? AI is the most notable recent innovation in automation. In May, AI leaders, including Turing Award winners and executives at OpenAI and Google, published an open letter warning of AI as a potential existential threat to humanity. Have they become modern day Luddites? The full text of their letter is: “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.” You can find the letter and its signatories at https://www.safe.ai/statement-on-ai-risk.
The statement is ominous, though it doesn’t detail the threat, perhaps because there is no consensus on how AI is a threat, only that it is. I don’t see the possibility of increased automation by AI as a threat; it’s other capabilities and applications that are concerning. It’s not a threat to humanity if AI is able to design the next generation of RF modules for smartphones. It’s also not a threat if the percentage of the U.S. population involved in farming is further reduced to 0.13%. AI can even write a decent “MicroBusiness” column [1]. It will cause disruptions for the people working in these areas, but if the pattern holds, other useful types of work will develop.
Were the writers of the open letter worried about AI taking over? Do we think that AI could turn our world into The Matrix? Dave was able to disable HAL. I suspect we’ll be able to turn off AI, if we want to. But could we become too dependent? As AI advances, there will be a risk of letting it make too many decisions for us. If we let AI determine which RF modules need to be designed, do we lose the ability to do that ourselves? There are far more concerning scenarios. Should AI write laws?
AI can do many things. It can create realistic images, videos, and audio tracks. We have already witnessed tensions from different groups embracing different realities and the resulting social turmoil. We’ve seen how social media has amplified those differences and helped encourage the embracing of different realities and even conspiracy theories. What happens when it is no longer possible to determine what is real and what isn’t? What happens when media platforms use AI to create and deliver customized content to each of us? That could be the most pressing near-term threat.
AI can do lots of things for us. Perhaps there are things we shouldn’t have it do. Do we want AI, or the way that AI is employed, to influence us? To motivate our thoughts and actions? Then AI might actually be in charge.
[1] F. Schindler, “Time to retire? [Micro Business] ,” IEEE Microw. Mag., vol. 24, no. 6, pp. 24–25, Jun. 2023, doi: 10.1109/MMM.2023.3255679.
Digital Object Identifier 10.1109/MMM.2023.3303588