I'm not going to dispute Ben Williamson's argument here. It is entirely plausible that much of the 'AI in Education' (AIED) research is, as he says, "junk science". What I will dispute the the suggestion - not made, but assumed, I guess - that this is new. While it is true that we are seeing "meta-analyses at an industrial rate and scale," this has been a trend for some time, certainly for as long as I have been a researcher, and the same case has been made - does the name Charles Ungerleider ring a bell? Did initiatives like the Campbell Collaboration solve the problem? No - here we are 23 years later, with the same old problems assessing the impact of a brand new technology, suffering from the same old delusion that we know what we want from education, as there is an all-seeing 'we' in that statement. Does a hammer produce better learning outcomes? That would depend very much on the discipline.
Today: Total: [Direct link] [Share]
