Stephen Downes

Knowledge, Learning, Community

This is a nice paper showing the history and nature of the association between artificial intelligence and the educational theory of constructionism and discussing some current and future prospects. Here's the short version of the history: " In (the) early days... there was a clear relationship between constructionism and symbolic artificial intelligence. More recently, this has changed with the wider shift in AI towards machine learning with neural networks and big data, which constructionism has mirrored." But something is missing, write the authors, in this new emphasis on neural networks. "With neural nets, students are now making ‘toy brains’. They are not programming at the level of concepts, symbols, and plans, but instead are relying upon crude approximations to how neurons work in brains." Maybe (and this would certainly reflect Minsky and Papert's criticisms of connectionism). But maybe programming at the level of concepts, symbols and plans lead students astray in precisely the wrong direction. No matter; the paper references numerous interesting projects and resources and should be read by anyone interested in the topic. The paper is open access but I found the BERA system made it difficult to download, so I saved a copy here: 13 page PDF.

[Direct link]


Stephen Downes Stephen Downes, Casselman, Canada

Creative Commons License.

Copyright 2021
Last Updated: Jun 18, 2021 3:29 p.m.