Dr Barbara Oakley

Using generative AI to strengthen and speed learning

About the event

AI can’t do your learning for you, but it can help teachers explain, engage, and generate opportunities to practise.

Unfortunately it seems that there are no quick fixes for acquiring deep expertise. There is nothing for it but continued practice to strengthen the neural connections in the brain and embed all the knowledge in your long-term memory.

However, as Dr. Barbara Oakley, Distinguished Professor of Engineering, Oakland University, explains in this fascinating and entertaining lecture, there are a number of techniques that can help you with that practice – and many of them can be enhanced by using AI.

The creator of Learning How to Learn: Powerful mental tools to help you master tough subjects, one of the world’s most popular MOOCs (Massive Open Online Courses), uses personal anecdotes, goofy animations and everyday analogies to explain what is going on in our brains as we learn – and, by extension, how AI works, and why it sometimes goes wrong. The ideas she discusses include:

Don’t just follow your passion: expand it

She explains how early specialisation led her to end up in a ‘career box’. ‘I’d done everything that everyone always taught me to: I’d followed my passion, I’d done something really well … instead of following my passion I should have broadened my passions.’ 

Understand the two learning systems in the brain

The working memory in the pre-frontal cortex holds only a limited amount of information, which needs to be transferred to the long-term memory in order to be retained. It can take one of two pathways: the declarative pathway via the hippocampus, in which you are consciously learning; or the automatic pathway via the basal ganglia, in which learning comes with repetition and practice. For example, she says, when you are driving a new route for the first time, you are alert and conscious of each turn that you make (declarative). Drive it daily and it becomes automatic, to the point that you feel that the car itself ‘knows’ the route, and you might accidentally take that route even if you intend to go somewhere else.

This unconscious learning is similar to what happens in large language models (the basis of AI), and it explains AI’s tendency towards hallucinations and confabulations because there is no conscious learning to direct its practice.

Use metaphors to teach – and AI can help

Metaphors (and Dr Oakley uses them so often it has probably become automatic for her) are not only good for explaining complex ideas in accessible ways but act as a bridge to learning (or explaining) other concepts. Generative AI can be useful in suggesting metaphors for use in teaching and can be prompted to come up with other ideas to stimulate curiosity and allow teachers to connect with students.

Although the education world as a whole is nervous about AI, Dr Oakley urges everyone just to let time go by. It is still new and exciting, and something that everyone wants to play with. However, she believes that the excessive interest in AI will decline, and we will be left with what is genuinely useful.

Watch the full lecture.