Disciples, Apostles, and Saints!
Since the public release of the large language models (LLM) which is what we tend to mean when speaking of AI — like ChatGPT and Claude — I have tried to hold an ambivalent attitude toward them, wanting to be both optimistic and pessimistic in appropriate measure.
There is little question that this tech can make our lives easier. They offer potential shortcuts to collating data, writing emails, and producing materials in seconds that otherwise take hours. This feels like a positive development. Except that the act of collating data, writing emails, and producing materials is actually important for our brains. This is when and how we actually learn.
Several years ago, Bill Allen described his method of note-taking to me. He taught his students to take notes in class by hand, then rewrite the notes at home, doing their best to remember what they covered in class. Then they have notes they can go back to a month later that will actually help them remember what they learned in class that day. What Bill was describing was a physical, labor-intensive process that leads to memory retention and concept learning. It follows the massive evidence that shows hand writing notes is dramatically better for learning than typing. Typing allows speed and easy correction, but the act of writing is where most of the knowledge acquisition happens! This means that we learn by doing the very tasks that many are seeking to have AI cut out of our lives.
And the evidence of the absence of this learning is starting to appear in studies, where students who use AI produce more than their peers, but less than their peers when AI is taken away. This means that there is a deeper ethical question about LLMs than whether to use them. It is what to make of the learning erosion on the human soul.
With love,
Drew+
