Thinking in Patterns — How Generative AI Supports Cognitive Schema Building

As part of my schoolwork, I watched a series on revolutions that changed history. We compared and contrasted the French, Russian, and American revolutions, with a quiz at the end of the lesson. I found it easier to retain information when I saw that all these revolutions shared similar causes, allowing me to see a pattern or structure behind them, instead of just memorizing hard facts. 

Our brain is wired to notice patterns. That’s why we can finish incomplete sentences or make assumptions. We call these schemas: mental frameworks based on patterns that the brain uses to organize new information. It helps us recognize relationships and easily understand complex topics. According to Piaget, schemas are building blocks of cognition, constantly changing and adapting according to where we are in our development. David Rumelhart expanded this into flexible networks, shaping our cognitive processes at any stage in life. Richard Anderson applies both in the context of education and learning, especially reading comprehension. These are theories that are now used in the development of generative AI as part of student learning.

In the classroom, strong schemas are crucial. Students learn more efficiently when they can recall information quickly, see the bigger picture, and apply knowledge to new or hypothetical situations. It’s employed when teachers use analogies of cake slices or pizza to explain proportions and percentages in maths. 

I’ve mentioned in multiple posts how AI models also learn through patterns. Essentially, generative AI has its own version of schema, analyzing millions of texts to understand how words and ideas relate. The AI process starts with embedding, where words and concepts are placed in a giant “map” with related ideas placed close to each other. Then, it clusters, similar to our cognitive process of categorizing. When prompted, AI will recall these patterns and apply them to the context it is given. Bear in mind that AI does not think, understand, or form meaning the way we do, and lacks the human ability to apply judgment through learned experiences.

So, what are some ways we could start applying generative AIs in the classroom to build better schemas? 

1. Brainstorming processesWhen I’m writing essays for my school assignments, I use AI to help me brainstorm ideas and organize my thoughts. With the image generation feature, I can request my AI tool to help me visualize a mind map, flow chart, or group my thoughts and interpret them into bullet point outlines with which I could write my pieces. It helps me see how my seemingly random ideas actually connect, and it’s very helpful when I need to see the big picture.

2. Using analogies to foster understandingSometimes, there are concepts in the classroom that I’ve not come across yet in my daily life. That makes it harder for me to understand and remember, because I don’t have my own preexisting schema. With such subjects, I can ask AI to explain them using analogies I would appreciate. For example, AI helped me understand DNA and gene expression by equating them to recipes and ingredients, and that mutations are “typos” in the recipe that change your resulting dish.

3. Checking and correctingAI can clarify any misconceptions when I prompt it to check my understanding or explain a topic differently. It can even preemptively check for mistakes by asking me questions to confirm comprehension. However, please note that AI can still make mistakes, so it’s important to double-check your facts through other sources.

4. Upgrade knowledgeThere are times I feel AI gives me an oversimplified explanation on certain topics, usually because I’ve prompted it to do so previously. When this happens, I can request that it to give me a higher level of explanation to mirror my level of understanding of the concept. I can also ask it to explain things in the context of where I am in my studies or which curriculum I am using at school. This helps me move from a shallow understanding to a deeper one and see it from a different perspective.

Teachers and educators can use AI to pre-teach background knowledge through AI-generated summaries or preview questions so students come to class prepared. They can also prompt students to use AI to make connections between different subjects and explore links between them, like putting History and English Literature together to see how the Industrial Revolution contributed to the significant rise of authors and journalists in 19th-century England. But remember, as helpful as AI can be, teachers must remain the central guiding post for all learning activities, ensuring that students still develop their own personal understanding and not blindly adopt whatever AI gives them.

Humans are naturally inclined to patterns. It’s in our DNA. Generative AI can help us see these patterns more clearly, spot relationships, and categorize accordingly. Combined with critical thinking, reflection, and active engagement, AI can make learning much more meaningful and rewarding than just memorizing facts.