Ed/Tech must-reads 170625

Hello must-readers, back again after taking an unannounced break for the long weekend last week.

GLAT: The generative AI literacy assessment test from Computers and Education: Artificial Intelligence

The need to develop GenAI literacy in educators and learners has been discussed widely but often in that hand-wavey “somebody should do something” way that is often mistaken for doing actual work. Jin, Martinez-Maldonado, Gasevic and Yan, take a step forward in this article to propose a format for evaluating this literacy with their GenAI Literacy Assessment Test, the GLAT. The paper indicates that results from this test predict performance with these tools better than self-assessment and it does handily provide the 25 question multi-choice quiz. There is something about MCQs that never really feels like a great indication of learning/knowledge to me - I understand the logistical value of them but being able to recognise a correct answer sometimes seems more like a test of knowing how MCQs work than actual understanding. Nonetheless, this is at least an applicable thing and worth looking over.

Michael Sankey: The portfolio as the place of representation for emerging professionals from Create.Share.Engage podcast

One of the other big “we should do something” themes in discussion of GenAI is the need for structural change to assessment to shift the focus of evaluation of evidence of learning from product to process. The use of programmatic assessment, where learning is tracked and feedback provided across the years in which students complete an entire degree (instead of within a series of unit based silos) has long seemed like a no-brainer for this. ePortfolios provide the container for a substantial part of this assessment. Catalyst’s Kristina Hoeppner talks through this with Prof Michael Sankey, who frames this as building the individual and building their home as a professional. I can appreciate that moving to a portfolio based approach can seem daunting, but it kind of boggles my mind that this isn’t far more widespread as practice, given the upheaval wrought by GenAI.

2025 Eportfolio Forum Call for Proposals from ePortfolios Australia

If you are already working with ePortfolios (in Australia), this longstanding annual event is the prime place to share your work and learn from like-minded colleagues. Scheduled for mid-November at the University of Tasmania (Launceston), the organisers are currently calling for short peer reviewed papers, lightning talks, discussions, workshops and posters. Submissions are due by July 14th

ChatGPT lost a chess game to an Atari 2600 from Extreme Tech

There was an exciting period (for some of us nerds) during the mid-90s when super-computers became competitive with Chessmasters, peaking with a match between IBM’s Deep Blue and Garry Kasparov. Deep Blue won one game, lost two and drew three in a series of six - but came back the next year to defeat the human champion by winning two games and drawing three. More recently, ChatGPT asked Citrix Engineer Robert Caruso to let it play against a chess computer which could only plan a couple of moves ahead to see how quickly it would win. This entertaining brief story details how well that went.