By Dr. Mike Walden
AI — short for “artificial intelligence”— is the technology wiz of the moment. One-third of businesses are now using AI, with the number expected to rapidly climb. AI programs can now be accessed online, with mobile phone versions to follow. Within a decade, almost everything will have been touched by AI.
But what exactly is AI? Some view it as a machine that can think like a human. Others see it as an easy way to find answers to questions. Some students see AI as a computer program that can write term papers for them!
The commonality is thinking, as implied by the word “intelligence.”
Many consider AI as a super-machine that can do thinking for us, and in many cases, that can do thinking we can’t do. So AI is a super-machine with superhuman qualities.
But is it? AI involves three components: significant amounts of data, computer programs to analyze the data and identify patterns and relationships, and conclusions based on analysis of the data. The term “artificial intelligence” was coined in the 1950s when mathematicians and statisticians began developing the concepts that would allow them to use large amounts of information to derive predictions. Only recently with the availability of large data sets and high-capacity and high-speed digital computers has AI become a reality.
The “machine-learning” of AI mimics human learning. Humans learn and make decisions by gathering information, looking for relationships and links in the information, and then making conclusions that help with decisions. The difference is that machines can do the same process much faster, more comprehensively and more frequently than humans.
Let’s give a quick example. A physician is examining a patient and attempting to decide if the patient’s symptoms are predicting a stroke. The physician would use knowledge that was learned in medical school about stroke symptoms, perhaps updated by recent articles the physician read in medical journals.
But using AI, the physician can feed the patient’s symptoms into a computer program, and let AI use access to millions of patient cases that are much more specific and current about symptoms specifically applied to the patient’s demographics. Additionally, the analysis can be done in minutes. The result is the physician has a much more comprehensive and reliable diagnosis for the patient.
The expectation is that AI will improve the quality of our lives, just like previous technological innovations did. I’ll never forget my late paternal grandmother telling me about the day the “lights came on,” as she termed it, meaning electricity had come to the farm. Not only did electricity provide her farmhouse with constant light, but it also allowed her to replace the old “ice box” with a refrigerator and to ultimately throw out the hand-powered washboard for a washing machine.
Most new innovations impact the labor market. Often new kinds of jobs are created, such as electricians when electricity was created and installed. But just as often some types of jobs are hurt by a new innovation. For example, the whale oil and candle industries experienced reduced sales when electricity became the common power source for lighting.
The adverse impact on segments of the labor force is what many worry about with AI, and there’s good reason for concern. AI-directed robots and machines will likely replace some jobs in warehousing, manufacturing, administration, and perhaps ultimately – even transportation. But the ability of AI to replace humans in some cognitive tasks may be the game-changer. With AI’s potential to examine data and information and make decisions means that jobs in areas like finance, accounting, banking and legal matters could be vulnerable to replacing humans with AI technology.
I have applied analysis completed by other economists to North Carolina and concluded that AI could eliminate almost 500,000 jobs in the state, approximately 10% of all jobs. The biggest losses would be in office support, retailing, manufacturing and food services.
Still, if economists are correct that our economy will grow faster as a result of AI, there should ultimately be new jobs available for those who lost their jobs because of AI.
But here’s the rub. First, the job losses will likely occur before new economic growth creates new jobs. So there could be a lag when the unemployment rate rises. Indeed, some economists predict the jobless rate could rise by several percentage points as AI is increasingly deployed. During this time period more resources would be needed to help the unemployed.
Yet even when the new jobs resulting from a larger economy are here, the jobs will likely require different skills than the skills the AI-created jobless currently have. As a result, North Carolina will need a massive retraining effort, likely directed at individuals who are much older than typical college students. Think of 500,000 individuals suddenly requiring new work skills. How will they be accommodated? Where and how will they learn? What will be the roles for public colleges and universities and private companies in meeting the educational need?
My conclusion is that AI is already here, and it will soon be everywhere. For the most part, I think AI will have positive impacts on our living, on our productivity and even on job creation over time. The challenge is the transition from our pre-AI life to our life with AI. During this transition there could be significant job losses, the need for support of the jobless and the task of retraining the jobless for new jobs. If we develop plans for these challenges now, maybe AI will ultimately be celebrated, just like my grandmother celebrated electricity.
But, you decide.
Mike Walden is a Reynolds Distinguished Professor Emeritus at North Carolina State University.
