If you’d rather listen to this newsletter, click here.
If you have been using the section on study methods in GRAD, you will have come across the name of Dr Barbara Oakley. She’s the one who failed Mathematics in high school and yet became a professor of Engineering after she upgraded her Maths using knowledge from brain science on how to learn. Her course on “Learning how to learn”, available for free on Coursera, is the most popular ever posted – it has had more than 3,8 million users since its launch in 2014.
She now has a new course, this time with Computer Scientist Dr Jules White, entitled “Accelerate Your Learning with ChatGPT”. One aspect they focus on is what Dr Oakley calls retrieval practice. She says that is the most effective learning strategy. To retrieve simply means to find and bring back something – you retrieve a ball from the neighbour’s yard, or information from the computer, or something you learnt yesterday from your memory.
When we learn, physical changes occur in the brain. The 86 billion neurons in our brains connect with each other in little clusters when we learn something new.
It is almost as if two neurons play footsy-footsy at the first connection opportunity, says Dr Oakley. There is a spark as neurons reach out and send signals. The signals jump the little gap between neurons, connecting them together. Our brains create sets of links when we learn something.
At first the connection is weak, just like on a first date. Nothing will come of it if it is not repeated. Connections strengthen with practice. The more we practice, the stronger the link between the neurons becomes, and the easier to recall when needed. On the other hand, if we don’t practice or revisit what we have learned, our brains have sweepers whose job it is to sweep unused connections away. That is why cramming is futile. You make lots of new connections and two, three days later you remember nothing. The sweepers have been at work.
Retrieval practice, says Dr Oakley, is far better than rereading, highlighting or concept mapping for effective learning. The other strategies have their place, but are not very effective if not combined with retrieval practice.
Retrieval practice simply means actively recalling key ideas from memory, strengthening those neural connections. AI is the perfect partner to help you do this.
First you can employ it as a kind of tutor. You could upload the document you are studying to your AI platform and ask it to “extract key ideas” to help you build a framework for your learning; or ask it to, “Act as an expert in breaking down (whatever the topic) and teaching through simple analogies and metaphors.”
Once you feel you have mastered that section of your work, you could ask AI to ask you questions, for example:
- “Give me 5 questions to deepen my understanding.”
- “Which 5 questions should I be able to answer to ensure that I understand the material?”
Jules White talks about something he calls a “Flipped Interaction Pattern”. It’s almost like a conversation with the AI. You could start with a simple prompt, like:
“Ask me questions one at a time to assess my knowledge of [the topic]” – “Ask me the first question.”
AI will, based on your answer, decide what to ask you next. The prompt’s basic pattern is: Ask me questions – One at a time – Start with the first question.
When we first talked about AI, we said you should use it to think more, not less. Coming up with better questions to test yourself on your knowledge and understanding is a much more effective way to learn than memorising answers. Get AI to help you do that, saying for example, “Whenever I ask a question, think of four more specific questions that, if you had the answer to, you could better answer the original question.”
We tried this. In response to the question, “What is the best way to make a bridge?”, chatGPT suggested these four more specific questions:
- What is the intended purpose and load capacity of the bridge?
- What is the span length the bridge needs to cover?
- What are the site-specific challenges?
- What is the budget available?
If this was part of your structural engineering studies, you would now know what to think about when planning a bridge.
You can also ask AI to help you formulate your questions, or prompts, when looking for information, saying something like, “Whenever I ask you a question, you will suggest a better version of my question that is clearer and more specific and ask me if you should answer that instead.”
We did this, asking, ”Who is the fastest person on earth?” The AI came back with, “Who currently holds the record for the fastest 100-meter sprint and what is their record time?”
Our next question: “Where should I enroll for my Honours degree?” AI’s recommendation: “What factors should I consider when choosing a university for post-graduate studies, and how can I determine which university would be the best fit for my interests and career goals?” Every time it added, as we asked it to, “Would you prefer me to answer this question instead?”
There is no right and wrong way to use AI, only more and less effective ways. Play around. See where it takes you.
Happy studying
The GRAD team
———————————————————
Click here to check out our other podcasts
If you like our content, forward and share!
GRAD – your guide to university success is a partnership project of Ruda Landman, StudyTrust, Van Schaik Publishers and Capitec Bank.