In a nutshell, " This paper (32 page PDF) addresses the question of how artificial intelligence should be designed and used to support learning rather than merely improve immediate outputs." The authors draw a sharp distinction between AI for work and AI for learning (illustrated) which I think may overstate the case. Still, the point stands. From there, they develop a three-part framework, based on (a) pedagogical foundation "to determine how generative systems can provide the precise support necessary," (b) an adaptive foundation, "organising adaptivity into a continuous four-stage cycle:" capture, model, adapt and evove, and (c) a responsible design foundation that "addresses how AI companions can act with integrity and uphold human values," based on security, transparency, accountability and inclusion. Via Philippa Hardman.
Today: Total: [Direct link] [Share]
