Great arguments backed by the data. I, too, am nervous about letting kids lose with LLMs in the classroom. There is little guidance and less research supporting it. Do you see any place for student use of AI? Coding? Feedback? Other forms of multi-modal generation? Given the way most students seem to be using the technology, I do fear that the adults who are skilled with using AI are projecting onto younger people a level of domain expertise they have built up over years and assuming the students will simply develop those as they use the LLMs. Great post and lots of useful data!
Yes, in some ways the digital native idea has it completely the wrong way round. In the early days of Google it was adults who had 20+ years of non-internet life & literacy skills who were able to easily find what they wanted with an internet search. A lot of kids were just clicking aimlessly not understanding what they were doing. Something similar is happening now with LLMs, with the added problem of hallucinations.
Great arguments backed by the data. I, too, am nervous about letting kids lose with LLMs in the classroom. There is little guidance and less research supporting it. Do you see any place for student use of AI? Coding? Feedback? Other forms of multi-modal generation? Given the way most students seem to be using the technology, I do fear that the adults who are skilled with using AI are projecting onto younger people a level of domain expertise they have built up over years and assuming the students will simply develop those as they use the LLMs. Great post and lots of useful data!
Yes, in some ways the digital native idea has it completely the wrong way round. In the early days of Google it was adults who had 20+ years of non-internet life & literacy skills who were able to easily find what they wanted with an internet search. A lot of kids were just clicking aimlessly not understanding what they were doing. Something similar is happening now with LLMs, with the added problem of hallucinations.