So I had a look at that paper getting all the press recently for showing a "cognitive debt" accumulated by students who use chatGPT only to write an essay. If you do a Google search for 'ai use results cognitive capacities' you'll find a number of studies making the same claim. The issue with this study and others like it is that they measure only for a specific set of capacities that were important for people who do not normally use the tool. They do not anticipate nor measure the development of new skills made possible through the use of the tool. I have other criticisms too, but the study has been stamped with the approval of MIT and Time Magazine.
Today: Total: [Direct link] [Share]
