Artificial intelligence (AI) software’s uncanny ability to emulate some important human behaviors has sparked speculation about its impact on employment. Will increasingly intelligent robots replace human workers, creating massive unemployment with heavy social costs?
Some writers, for example, The New Yorker’s John Cassidy, recall the textile industry’s tumultuous history in the early 19th-century United Kingdom, when sophisticated new automated equipment replaced then-highly skilled production workers.
Many displaced workers became violent, including through attacks on the job-eliminating machines. Known as ”Luddites” after a leader of these activities, these rebels remain etched in the annals of industrial history for the ill social consequences caused by technology-driven unemployment.
Fast forward to the present, John Cassidy asks if we can do better as AI job loss looms. To be sure, job losses in some activities are real. In fact, AI-driven employment changes are emerging in the writing of software code.
Google and others are cutting by 20% the labor needed to write code by replacing low-skilled coders in the programming process with AI. This kind of productivity improvement is nothing new in software development.
I recall early years in digital system development at RCA, when integrated circuit chips had several hundred active transistor devices that were laboriously interconnected with simple design rules to perform desired data processing functions.
The process was so laborious that the floor of a room was used in the design process, with engineers writing the interconnections on blueprints. At RCA, we designed some avionic systems with such labor-intensive methods.
This did not last long, as software was designed to automate the process. We soon built chips with thousands and then millions of interconnected transistor elements.
So what happened to the engineers who were replaced by the improving software? Some moved into chip design software development, and some moved into system design.
I don’t recall that we had an employment problem because the company continuously invested in training the engineering force to allow people to move into different jobs as the technology improved.
Today, chips with billions of transistors are designed with highly sophisticated equipment. There is a global shortage of chip designers capable of managing the process.
That said, there will be employment dislocations as AI applications move into many business functions. Repetitive business operations are prime targets for AI implementations.
Employees with such functions will, in fact, be replaced and need to be trained for jobs requiring new skills. But here again, we have seen such a job evolution process before.
What happened to the many thousands of people who used to perform clerical functions or operations, such as telephone switchboard operators? New jobs emerged as technology matured.
The simple but important answer is that technology can replace functions performed by humans only up to a point. In the end, human intelligence underwrites AI systems. The key to avoidance of technology-driven unemployment is national education that trains a technology-ready workforce.
Henry Kressel is a technologist, inventor, author, and long-term private equity investor in technology companies. An industrial pioneer, he headed electronic device research at the RCA Laboratories, where he introduced the modern electro-optic technology.
