By Dennis Sale
THOSE who have followed my columns will know that I have written previously on the potential impact of artificial intelligence on teaching and learning. However, the field is rapidly changing almost weekly and even the experts (and there are many of them) are unsure of what will happen specifically, how, and when.
What we do know is that with more data and computational power AIs are becoming increasingly capable of predicting almost anything in complex environments if extensively trained in that knowledge field. As high levels of prediction are the outcomes of expert thinking skills, this means that future AI tools will have a heightened intelligence – what is referred to as artificial general intelligence – that is as good or better than that of humans.
While the human brain has massive neural capacity, it also has significant systemic limitations. Firstly, all new learning must initially pass through and be processed in working memory which has limited capability (eg between five to nine bits approximately) – that’s why learning a new language takes ages.
Secondly, the transmission of information from one human brain to another involves the teaching and learning process – as it simply cannot be immediately and effectively downloaded from teacher to learner. The typical school student will spend around 15,000 hours at school and experience, as well as cognitive science, supports the notion that what is taught in this long learning episode is not well processed and retained over time.
Anyway – and this is the kicker – AI machines, unlike biological brains, can store information in a much denser scalable way and simply transmit this digitally and instantly to each other. That is how GPT4 knows thousands of things more than any one person. Furthermore, as technology advances, we can expect storage capacities to continue growing exponentially, enabling much faster access and retrieval of information.
A potential outcome of AI development is that the machines will vastly surpass human intelligence, including problem-solving, creativity, and learning. To illustrate, we know that humans who are expert in an area can assimilate new knowledge much better and quicker in their field than non-experts (because they have a solid comprehensive base already established in their long-term memory). Now, imagine AIs which can access all the available knowledge in the world and can process it at incredible speeds; that’s a transformational learning capability. The big question then is how this will transform the future of human learning?
Impact on the educational landscape
One of the major aims and challenges of educational systems has been to provide a differentiated curriculum that enables instruction to be fully tailored to meeting the learning needs of diverse student groups. While existing technologies have provided a wider range of differentiation than in the days of “chalk and talk” teaching, it is still a major educational challenge. However, new AI technologies are creating a revolution in this area. The major impacting features include two main interrelated determiners of effective learning and self-regulation:
1. Customised personal curriculum
As a learner you will have your own AI agent that will fully customise the curriculum personalised to your existing level of understanding in all subjects and learning goals, available 24/7. It is not subject to tiredness or mood changes, always patient, and even predicting your learning needs through processing what and how you have been learning and where you are having difficulties. It can also show empathy through active listening techniques such as “Is that what you’re saying”; “What do you think might help here”; and even verbal affirmations like “I understand what you mean”, etc.
Central to effective learning is the importance of receiving ongoing, clear, and focused feedback on one’s learning performances. Your AI agent will be right on this task as new models will capture all the interactions you have with them. Hence, it will be able to model your learning progress, what you have learned effectively and where gaps and challenges exist and, most importantly, quickly develop effective instructional strategies to keep you on track.
2. Learning coach
I have written extensively on the importance of metacognition (eg Sale, 2020), and why this capability is the most important 21st century competence. Metacognition is our uniquely human capability to reflect on our own thinking emotions beliefs and behaviour. People with high metacognitive capability can plan, manage, monitor, and evaluate their own learning and behaviour better than those lacking such ability. This enables them to self-regulate more effectively, which includes emotional management, showing resilience under pressure, and maintaining self-efficacy and well-being. As Noushad (2018) concluded:
“…metacognitive strategies are essential for the 21st century because they will enable students to successfully cope with new situations and the challenges of lifelong learning.”
However, schools typically focus mainly on teaching subject content knowledge. They also provide other learning and wellbeing support. However, they also have limited resources and teachers are very busy people. Even though metacognitive skills have a high impact on student attainment, the situation is still largely as Ambrose (2010) notes:
“Unfortunately, these metacognitive skills tend to fall outside the content area of most courses, and, consequently, they are often neglected in instruction.”
It is here that your AI agent can help you to develop metacognitive capability and maintain effective self-regulation as its always available, and it will get to “know you”. The latter could be scary if you are influenced by films such as Ex-Machina, where a male researcher is tasked to identify if a female AI can pass the Turing Test (this is a test of a machine’s ability to exhibit intelligent behaviour equivalent to, or indistinguishable from that of a human). He develops a friendship with the AI and helps her escape the laboratory, but it does not end well for him.
However, I am more optimistic about AI and am more worried about the military rather than interpersonal consequences (though I don’t pretend to be right). Hence, with its ever-advancing capability for high-level reasoning it can be your critical friend, ensuring that you’re thinking and self-regulating is progressing as planned, keeping you sharp, focused, encouraged – maybe even entertained.
In summary, AI agents can become a powerful and enduring personal effectiveness resource for all of us, and especially for students. Firstly, in terms of pedagogy (eg teaching and learning methods) it will draw on the most current knowledge on how humans learn from cognitive neuroscience and create a customised curriculum and instructional system for meeting your learning goals. Secondly, and equally important, it can be an informed mentor and coach, enabling the acquisition and application of effective self-regulatory skills essential for personal growth and wellbeing.
-
Dennis Sale worked in the Singapore education system for 25 years as adviser, researcher and examiner. He coached over 15,000 teaching professionals and provided 100+ consultancies in the Asian region. Dennis is author of the books Creative Teachers: Self-directed Learners (Springer 2020) and Creative Teaching: An Evidence-Based Approach (Springer, 2015). To contact Dennis, visit dennissale.com.
