Germane cognitive load
That seems to be what's missing in how some people use AI.
And that's what makes them dumber.
A recent MIT study found that using LLMs (large language models) – or types of AI programs that run ChatGPT, Gemini, and other chatbots – can impair cognitive abilities, including memory and critical thinking.
The study involved individuals who were asked to write essays using either their brains, search engines, or ChatGPT.
Using ChatGPT resulted in bland, unoriginal, generic, and characterless essays. Much like how robots would do it, I guess. Too perfect. To the point of being boring.
Asked to redo previous essays, the ChatGPT group remembered very little of what they wrote about.
Dr. Poppy Crum, adjunct professor at Stanford who specializes in neuroplasticity and the use of technology, said it made sense when viewed through the lens of cognitive load theory.
Cognitive load theory assumes that our brains have limited resources when processing information. It is about the processing burden of thinking and learning.
In a podcast with Dr. Andrew Huberman, Dr. Crum talked about the three types of cognitive load: intrinsic cognitive load, extrinsic cognitive load, and germane cognitive load.
Intrinsic cognitive load refers to the inherent difficulty of learning something, say, calculus. Extrinsic cognitive load is about how the material is presented, or the challenges presented by the environment. Germane cognitive load is about the mental effort to learn something. It is about building mental schemas or representations that make learning possible.
It's hard work. But it makes us smarter. It helps us remember.
The problem is, some people outsource the hard work of thinking instead. Is that the best use of AI for us humans?
Before we lose any more neurons, we need to have a better appreciation of germane cognitive load. Maybe there's a prompt for making that happen.