Is learning finally obsolete?
The power of AI driven technology to impact education drastically was recently shown by ChatGPT. Overall, the discussion around what AI could likely do has swung between overly pessimistic or overly optimistic. In the teacher world, it has tended more towards fear and disbelief as students could simply ask ChatGPT for an answer. It has made more than one teacher ask, “What is the point of all this?” or “Why do I even try?”. These are very legitimate feelings. For example, why should one even attempt to teach essay writing when a computer programme can not only do it better than any student but can also do the favour of dumbing it down for the student so the teacher would think it is in fact the student’s own essay.
So, is learning finally obsolete? Have we finally arrived at the point in human history where learning can finally stop? To answer this, let us explore a certain classification of cognitive complexity developed by this smart dude named Benjamin Bloom in 1950s and has then been revised over the years. It is called Bloom’s Taxonomy (1).
How does learning work?
Bloom’s Taxonomy nicely arranges cognitive skills and their difficulty level, almost like different levels in a computer game. It shows how knowledge can be arranged from the lowest order, “Remember”, moving through levels “Understand”, “Apply”, “Analyse” and “Evaluate” to finally reach the level “Create”. The figure below shows these classifications along with a description. Importantly, each level builds on top of the other. You cannot analyse a problem, let alone create something new if you don’t remember the basic facts and understand what the different concepts mean. Why is this important?
The world is becoming ever more complex and the problems that we face are also getting more complex and intertwined (80000 hours has composed a list of these pressing problems). One can’t hope to solve challenges like global warming or pandemic risks simply by remembering the structure of a carbon dioxide molecule or simply understanding how viruses transmit. So, there is no question why we need citizens who can think critically, analyse complex problems and work together to create out of the box solutions to these multidimensional problems. Considering this, it is only fair to demand educational systems to train and produce students who can think critically and help solve world’s most pressing problems. This changing role of education was in fact clear since the dawn of the internet era when knowledge became more available and remembering facts was not the cool thing anymore.
So, do we stop teaching facts and basic skills?
We need our students to develop into creative thinkers. Does this mean we should stop teaching facts or stop testing understanding of simple concepts, especially because Google or an AI model already knows the answer? If you had asked me a few years ago, before even AI came to shake our worlds, my answer would have been a very passionate yes. This would have likely followed by a mini lecture on how it is pointless to learn facts as we have all the information in the world in our pockets, thanks to our smart phones. This idea was so deep rooted, that I even secretly took pride in how I will only aim to teach “deep understanding” (whatever that meant!) and not facts. In classroom practice it meant that, I threw difficult “self-directed inquiry based learning” tasks at my students hoping that through analysis within student groups they would not just understand the topic but also remember the facts (but who needs them anyway, I thought). As a bonus, they learn to work in groups, a crucial skill in this modern world. At least three birds in one shot, I thought smugly. This is also what I got wrong. I assumed that my students are like members of a team with multiple areas of expertise working together to solve a challenging problem. I assumed learning new information looks the same as using existing information. I mistook student engagement (and sometimes frustration) for learning.
But according to wise Mr. Bloom skills are built bottom to top. You cannot effectively learn skills that involve higher order thinking without having factual knowledge in your long-term memory. Tasks involving evaluation and analysis rely heavily on having solid chucks of lower order background knowledge and skills (understand and remember). Examples: You cannot analyse World War II if you do not remember what countries were involved in the war. You cannot understand why the sky is blue if you do not remember that light is a wave. You cannot evaluate the player moments in a football game and make decisive passes if you are still learning how to kick a ball. Complex knowledge and skills are built on top of simple knowledge and skills.
What is the take-away message?
As Daisy Christodoulou says in her blog: it doesn’t matter if we set our students tasks that can be easily solved by computers. It doesn’t matter if they produce writing that is weaker than that of ChatGPT. The easy problems and the weak writing are milestones on their journey to mastery which cannot be skipped or outsourced.
This realisation of the importance of fundamentals and building background knowledge has lifted considerable weight off my shoulders as a teacher. Just little like Daniel in The Karate Kid is forced to repeat “wax on, wax off” to hone his reflexes, teachers can and should spend time teaching students things that computers already know the answer to or can do way better than them.
In fact, Scott Young thinks that the value of old-school book learning and having deep knowledge in your head would matter more. He argues: “LLMs [Large Language Models] may accelerate the need for knowledge, since the combination of [knowledgeable person + ChatGPT] will outperform [ignorant person + ChatGPT] by a wider margin than exists between the two people on their own.” For example, you could create a table faster and better if you know how to use a circular saw compared to someone who does not.
In short, if we want our students to grow into citizens who can solve complex world problems, we must make them write those essays. Having knowledge in our head is no substitute to looking it up or asking an AI language model. So, no, learning is not obsolete.
(1) Krathwohl, David R. "A revision of Bloom's taxonomy: An overview." Theory into practice 41.4 (2002): 212-218.
