Artificial Intelligence (AI) is changing learning
Imagine a learning system that adapts to you personally. If you get a test question wrong, for example, it suggests you look at the relevant content afresh then tests you again.
But, I hear you say, there are plenty of systems doing this already!
OK, then. Imagine a system that notes that you got pretty close to the right answer and explains at which point you went wrong.
Not many systems are doing that, and where they are achieving this sort of feedback, it’s for very specific and precise errors.
Beyond that point, the only limit to feedback and adaptation of learning is one’s imagination. Putting a vision into practice is then the task of those working on this year’s hot topic: artificial intelligence (AI).
But what is AI? Definitions are ten a penny. Let’s just say that my electric kettle knows when to turn itself off, but I wouldn’t count that as artificial intelligence. But an algorithm that detects a helpline caller’s tone of irritation and apologises certainly is using artificial intelligence.
AI might be this year’s hot topic, but people have been suggesting how it might help for many decades. An important recent article by Woolf and others raises four challenges for those interested in learning. The first challenge suggests that every learner should have an AI mentor. Such a mentor would observe the learner and deduce how he or she was getting on by looking at behaviour, test answers, emotional responses and so on. To put it in basic terms, everything about the learner would be fed through an algorithm and the content would be adapted accordingly.
Or, as the authors put it, “AI provides the tools to build computational models of students’ skills and scaffold learning. Further, AI methods can act as catalysts in learning environments to provide knowledge about the domain, student, and teaching strategies through integration of cognitive and emotional modeling, knowledge representation, reasoning, natural language question answering, and machine-learning methods. Such tools provide flexible and adaptive feedback to students, enabling content to be customized to fit personal needs and abilities.”
London’s Institute of Education sees AI as a possible replacement for exams, testing, and assessment of all kinds. In a recent paper called Intelligence Unleashed, Professor Rose Luckin imagines a world in which AI can follow a learner’s progress continuously and provide appropriate support when needed. This, she believes, would do away with the stress attached to exams and the need for “stop and test”.
“This new approach, which aims to support both teachers and learners, would not attempt to replace the high quality human interactions that teachers bring to the learning environment. Nonetheless, they could still have a useful place within the education system.
In practice, AI technologies would build a model of the curriculum knowledge and trace the learner’s performance against the model’s knowledge. This system could be applied to SATs exams, GCSEs and A-Levels."
Obviously, learning platform designers such as Coracle are already thinking about algorithms that take learning in this direction. As a result, you can expect to see the development of more flexible and responsive courses in coming years. They will likely be friendlier and aimed less at the traditional exam format, and learners will (hopefully) be more engaged as a result.