Ed & Tech must-reads 4th August 2026

The missing piece of higher education's AI response: Adaptive capabilities from Needed Now in Learning and Teaching

More and more, I’m struggling to see how we get there when it comes to the use of GenAI in learning and teaching. Wherever “there” is meant to be. To be fair, educators, educator advisors, and institutions are trying our best to support learners in a time of radical change. (Or at least what we are constantly told is a time of radical change). This well-written and thoughtful post from Paula de Barba (Monash U) summarises a larger report for Australian government HE regulator TEQSA about what is needed to assure learning with regard to “adaptive capabilities”. But the heart of what I’m reading is what I have heard in the 20+ years that I have worked in the TEL space - we need to bolster students’ digital literacy in terms of their ability to use tech and critically process information. If we are still saying this now, why or when will things be different?

Much is made of the need to adapt to an epistemic space where knowledge is claimed to now be co-created by people and AI. In the exemplars provided though, it feels a lot more like just being richer web searching with a natural language interface. An example put forward as hybrid metacognition looks uncannily like tweaking a search term to get a more specific answer. Is that all we are actually doing? Clearly I am leaving out an important part in this musing - that of how we minimise or eliminate cognitive surrender to AI, particularly with people without the existing body of knowledge to know whether what is output is worthwhile or not. And maybe it is still possible to build agents to conduct genuine Socratic inquiry which provides individual support to student learning which is not otherwise possible. But overall, the more I read from the absolute torrent of writing about GenAI in education, the less value I am seeing, and when stacked up next to the mountains of educational and social harm that this industry is bringing, I have to ask why we persist so uncritically. Sorry, that was a bit heavy for a Tuesday.

Adults have struggled to set rules for AI in school. These teens figured it out from NPR

Until they put me in charge of things though, we live in the GenAI rife world that we inhabit. This post from the US National Public Radio network focusing on K-12 education describes a very ‘students-as-partners’ initiative intended to develop a policy framework for the use of GenAI in education. Key items are providing sufficient development of digital/AI literacy in students, banning its use in graded tests and supporting oral assessments as a fallback when cheating is suspected. Students took a harder line on directly inputting assignment prompts into AI tools than most institutions but seemed to allow it for editing, brainstorming and studying after 8th grade. (I think this speaks to a certain respect for mastery and excluding GenAI until a nominated level of skill has been attained). They also took a hardline on student’s using GenAI to bully others, recommending suspension or expulsion. It’s an interesting read the full text of the students’ ‘legislation’ can be found at the end of the post.

Around the LMS Grounds with Phil Hill (and Moodle)

It is evidently big LMS conference season in the US and Phil Hill from the On EdTech blog has provided a useful wrap up of key points of discussion from the Blackboard, Canvas (Instructure), and BrightSpace (D2L) events.

Blackboard, coming off some massive corporate restructures and the brink of collapse, worked hard to project hope for the future without much in the way of new functionality beyond a refreshed gradebook interface (nothing to sneeze at) and an updated content editor. They also launched their new purchase, Cursive, an academic integrity detector focused more on process than product.

Instructure had their own confidence building to do after the massive Canvas hack a couple of months back, so security and privacy functionality took centre stage. This meant that less time was made for discussion of new features, with a heavy emphasis on AI

D2L (Brightspace) also went hard on the AI side of things and evidently took some time to have a poke or two about the Canvas breach. There also appears to have been more time for discussion of useful but unsexy nuts and bolts functionality.

To round out the group, Moodle has rebranded and relaunched their (admittedly rich) open collection of community generated plugins in the “new” Th. Nice to see that Moodle Poodll is still around and kicking.