Education in the time of AI, or The evolution of education
Depending on whom one asks, the purpose of education differs. I’m partial to two explanations: that it’s intended to equip its students with the capacity for rational and independent thought, or to turn freethinking individuals into gears of the machine that is society. Though these seem to be contradictory, as technology and society progress, they are converging, and both have held true throughout the history of widespread education. The advent of artificial intelligence chatbots such as ChatGPT have caused controversy among many domains, and I’d like to weigh in with my thoughts on education.
The two main controversies arising from ChatGPT in education differ greatly. The first is a knee-jerk reaction to students who wish to cheat. ChatGPT can write compositions on anything and everything, and is already beginning to see use up to the post-secondary level. Though at that level, it isn’t quite as eloquent as a diligent student, it certainly writes well enough to merit a passing grade. This means that cheating for homework assignments is now possible across all the liberal arts, and even some of the scientific ones, where research papers need to be written. Formerly, cheating was limited to more quantitative areas of study, where proofs or right answers would be copied, but AI chatbots can contribute to even these, through plain-language input and beautifully formulated output.
The second concern that the advent of ChatGPT introduces is existential in nature, the very point of education itself. Previously, coherent writing was strictly a human ability, and one that required a lifetime of practice. Now, there seems to be a ghost in the machine, an AI powerful enough to fool humans with its output. Do students still need to learn to read and write if chatbots can synthesize and generate better than a human? More than that, what skills should students learn to stay relevant and abreast, if not ahead of the curve, to prepare themselves for a future where the majority of jobs can seemingly be automated? What hath God wrought?
I find that the two issues that AI introduces to education, like the two reasons for education, converge as well. History provides much of the evidence behind the second reason: originating from the religious state in the eighteenth century, from the capitalist state in the nineteenth, and then all other states in the twentieth. The need to turn children into obedient Christians, then cogs on the assembly line, is stamped all over education, from discipline and obedience to alignment of class schedules with parent workdays. As the nature of employment has changed, so has the education provided. From religion to rationalism, cursive to typing, and arithmetic to math, curriculum has aspired to provide students with the tools they need to succeed: the advent of keyboards supplanted a need for a fast-shorthand, the introduction of calculators caused less emphasis to be placed on Kumon-style calculation. Thusly, with the improvement of technology comes the increased focus on thinking and creativity; but today, that is what it takes to be a cog in the machine.
Back to the concerns that chatbots have raised in education. Homework has always been created in the service of learning. Shakespeare and calculus are taught in school not because the knowledge or ability in itself are useful, but rather because they are conducive to analysis and logical thinking. The primary question then, is what is to be taught. Just as calculators and Wikipedia have reduced the importance of long division and rote memorization, AI will certainly alleviate some mental burdens to the benefit of others. The issue, is that it’s hard to discern exactly what that is yet.
Plato believed writing was a crutch that weakened the mind, and the truly wise could contain everything they needed to between their ears. Though he probably consisted of the entirety of human knowledge at that point, the same argument could be said of calculators, the old “you won’t always have a calculator handy in your pocket”. So, making similar comments about chatbots is a surely doomed endeavour. This isn’t to say that writing, or anything else an AI can do, for that matter, is already outdated and unimportant to learn, but there will certainly be effects on what is considered important. The issue here, is that the output of ChatGPT emulates much of what we humans deem to be basic thinking: recitation of information and basic logical deduction and inference.
For virtually all of history until now, creation has always defined humanity. Some animals make shelters, but aside from that, the implementation of knowledge, logic, and creativity to make something new has been what has spurred human progress, and resulted in names being enshrinned in the annals of history. So, to have a piece of machinery read through all of that, and then use it to draw conclusions and produce something new is not only extraordinarily impressive, it’s also threatening. If, as AI researchers predict, the technology will only continue to grow at an increasing speed, it won’t be long before the output of a single generative AI exceeds that of all of humanity. But before that, or even despite that, writing in education will have to persist because it is the only signifier of human thinking. Just as Shakespeare and calculus before it, even if chatbots will be used to create virtually all the textual output that one needs in their adult lives, writing is still an exercise in ideation and thought structuring. Until technology is able to pluck the very threads of thought from our heads, and then restructure them into something meaningful, writing is not only the best way to communicate our thoughts, but also the best way to practice thinking as well.
As chatbots effectively turn convert written logic into a query, and answer, and then back, perhaps learning to communicate effectively with AI is the skill that has to be learned; machine whisperers carefully formulating expressions inputs to get the best possible output. But that defeats the point of chatbots, whose utility consists of standard input that anyone can write. If the growth of AI doesn’t prove to be the meteoric ascent that everyone has predicted, maybe it will be humanity training in subservience of these AI, seeking to compete at filling the gaps. For now, there is nothing we can do but assume that AI will continue to act in the way it does, large language models approximating conversation and statistical analysis drawing correlations, leaving thinking to the humans.
Whatever AI is eventually capable of, chatbots have shown us that it will become eventually impossible to find jobs in copywriting. Chatbots are not only already able of comprehending conversational input to create well-written output, but are also capable of writing in different styles: Tupac in the style of Shakespeare, or Moliere in the style of Biggie. Despite this, writing is still an important skill to learn at school, as it trains creativity, structure, and logic. In the unlikely situation where writing becomes obsolete in the future, the skills garnered by writing will be just as important as the ones from doing calculus or reading Shakespeare. There will always be cheaters in school, but for those who won’t deign to, I say rest assure that no time is wasting doing the mental weightlifting that is writing.
