Ben Williamson writes, and I agree, "The underlying problem is that there is current desperation to show the causal 'effects' of AI in education - whether good or bad - and this is leading to a rush of studies that immediately gather huge public and media attention despite their significant methodological shortcomings and limitations." I'm less inclined to blame "low-quality peer review and high-speed editorial and publishing processes" - both good and bad research can be found anywhere (and Williamson himself cites an ArXiv preprint along with some Hechinger Report articles). The real problem, as Williamson recognizes, is that such studies try to isolate a single factor - AI - in what is a complex process where no single cause explains anything. Academics and journalists alike should do better.
Today: Total: [Direct link] [Share]
