Luddites and philosophers
It’s interesting to think about how people imagined the future half a century or a century ago, with robots everywhere taking over human jobs. Isaac Asimov built an entire universe around the story of robots, yet even three years ago, it was the dancing robots of Boston Dynamics that captured our imagination. A kind of Luddite sentiment emerged, a fear that these robots would replace the lowest-paid and therefore most vulnerable workers. According to these visions, the new proletariat would resemble not the proletarians of the Modern Era, but those of the Ancient Era — urban masses whose only real assets were their children. Many suggested that the spread of robotization should be limited to protect the disadvantaged.
Artificial intelligence-enhanced chatbots then radically changed this forecast. It seems that it won’t be the physical workers who are replaced first by new technologies. This isn’t entirely new; Hans Moravec noticed a paradox in 1985, which was named after him: it’s relatively easy to get a computer to perform at adult levels on an intelligence test or play checkers, but difficult, or impossible, to match the capabilities of a one-year-old in terms of perception or movement. The issue is not just the difficulty of these tasks (although Steven Pinker’s 1994 observation that the hardest tasks are easy and the easy tasks are hard is instructive), but also the cost. While developing and training Chat-GPT might have cost nearly four million dollars and its operation isn’t cheap, the cost per user isn’t much. In contrast, an intelligent robot requires motors, sensors, and electronics for each unit, and it will be a long while before this is cheaper than human labor. However, there are job roles that can be cheaply and efficiently replaced by artificial intelligence, such as customer service agents, claims adjusters, accountants, marketers, and legal assistants. The transformation of the labor market occurred sooner than expected, not affecting the lowest social status jobs requiring the least education, but rather those intellectual tasks that can be performed with just a high school or bachelor’s degree. Masons, electricians, and even retail stock fillers will still be needed for a long time, but it’s the middle ranks of professions that will disappear. This could lead to greater social division than ever before, as the spectrum is not continuous and jumping from one end to the other requires a significant leap. What must schools do to ensure their students are on the winning side of these changes?
It may not be surprising that we might find answers to this question in ancient Greek philosophy. Since Socrates, we’ve known that the real achievement isn’t in providing an answer to a question, but in asking good questions. Socrates himself often just asked questions in his dialogues, and his interlocutors would reach the truth by responding to them. Using artificial intelligence also makes it clear that the most important thing is how well we can formulate our questions and requests to the program. It’s no coincidence that the role of the prompt engineer, a specialist in formulating questions and instructions for artificial intelligence, has become the newest profession that nobody even imagined would exist a few years ago.
Classical education is shockingly answer-focused; consider that a student spends most of their educational career receiving prepared answers from teachers or materials, then during assessments, they must provide the answers to set questions. The art of questioning isn’t emphasized in learning. I sometimes conduct what I call a “reverse test,” where I provide the answers and students must determine the questions. It’s a much harder task than it seems at first, and students often get upset by it. The essence of question-focused teaching is whether a student can pose valid questions, which requires at least as much knowledge as needed to answer them, if not more. Discovery-based math education approaches this idea. To ensure that our students are proficient users of artificial intelligence tools, they certainly need to be capable of questioning.
Furthermore, basic skills that classical competency-based teaching emphasizes come to the forefront. Foremost among these is reading comprehension, which is essential for understanding and appreciating the responses generated by artificial intelligence. Teaching advanced reading comprehension skills is also important for effective use of artificial intelligence, as it requires the ability to analyze, evaluate, and synthesize texts. Developing these basic skills (reading comprehension, mathematical reasoning, scientific thinking) cannot just be a well-sounding prelude in a fact-crammed curriculum if we want our students to stand their ground in the world.
To handle artificial intelligence effectively, it’s crucial to understand how the system thinks. Since these programs are trained on texts and images created by humans, it’s not surprising that we best understand their mode of thinking by knowing how humans think. The field concerned with this is called philosophy. We can expect that philosophers (alongside linguists and cognitive scientists) will be in higher demand, and that philosophy, once considered a navel-gazing hobby by many, will regain its status as the queen of sciences, as in the days of Socrates. It would be wise to quickly figure out what kind of school can teach thinking and prepare children for this brave new world.
