Abstract

ChatGPT and other generative models raise a problem for the use of papers and dissertations as a way to assess students’ abilities. Is this just a manifestation of a more general problem, that is, the tendency of our education systems to raise “stochastic parrotsµ? I suggest an affirmative answer, but with the qualifications that human understanding may occur through different cognitive strategies and that stochastic parrots exhibit, after all, one of those strategies – although a shallow one.