A potential unanticipated consequence of AI is a hollowing out of the workforce as entry-level white-collar jobs are eliminated, and as the perceived value of higher education diminishes.
The effectiveness of AI is its ability to increase productivity by eliminating the need for workers to complete structured tasks (e.g., writing code, financial analysis, and even medical diagnosis). But while the sophistication of these models is increasing rapidly, they still depend on a person being able to conceptualize problems and so ask the “right questions”. Currently, many such workers exist and so the immediate impact of AI on productivity will be significant.
However, this success may be temporal. Take as an example the role of data analyst. The effective analyst knows from experience what information is useful, what is extraneous, and how best to organize that information for decision making. The production of that information is then generally relegated to entry-level analysts. Then over time, these “junior” analysts acquire the experience needed to move into the more senior role.
Today AI can produce the same work as these entry-level analysts in less time and at no appreciable cost. As companies learn to exploit this cost-saving opportunity, the number of such entry level jobs will decrease. But then what happens as the more experienced workers leave the labor force and there are fewer then-seasoned workers to follow them? Without someone to ask the right questions, the value of AI might diminish.
AI is very good at answering questions and even anticipating follow up questions. But lacking “values”, it (for now) is much less good at knowing what questions most need to be answered.
This brings me to my other related concern. Historically, the role of higher education has been the transfer of knowledge and the development of thinking skills (particularly critical thinking) and ability to communicate effectively. But already the substantial cost of higher education paired with the reduction of well-paying entry level is changing the calculus of young people. In 2016 approximately 70% of high school graduates went to college. In 2021, the number was down to 62%. So, not only might there be fewer knowledge-based entry level jobs offered, but there will also be fewer qualified workers for such jobs.
There of course will continue to be those who do have the knowledge and ability to effectively exploit this technology. And so, economic progress will likely continue. But a concern is that the rewards will accrue to a shrinking number while most will find themselves in less well-paid, non-knowledge jobs with limited opportunity for advancement. This begs the question: Can higher education adapt to help bridge the experience gap between younger and more experienced workers? Or should we instead adopt an alternative model where most pursue vocational education directed at working effectively with existing AI models while a smaller number to pursue the riskier but potentially more rewarding traditional route?