Stephen Downes

The focus here is on U.S. schools, but there is a broader point worth making. " Too many of our schools sort children more reliably than they develop them," writes Elena Silva, referencing "segregation by race and income". This is a problem because "Learning is fundamentally relational. The brain develops through attachment and interaction." This sort of interaction works best with diversity, not homogeneity. Sorting, in other words, breaks interaction. "This relationship-centered approach is not a soft alternative to academic rigor; it is its essential mechanism."

This is interesting, though the full report is behind a stupid subscription wall, which is very frustrating and makes me think a lot less of them. The premise is that AI is interfering with student collaboration because all the students depend on the same AI model, which means when they come together they're already saying the same thing. Apparently the report documents this in more detail, though we get a good sense of it on this open page. The tease is 'seven methods that drive learning in collaborative groups'. We are led to believe the DEC report details a solution, though this is craftily hidden - and thus, to my view, is nothing but a fake. Here also is the press release.

According to this article, "Recent advances in AI raise the possibility of building something education has never really had: a virtual laboratory for learning. At the center of this possibility are synthetic students." The problem, of course, is that we don't know how to simulate students. "When you ask a capable model to play a struggling student, its mistakes tend to come out inconsistent or superficial."Now are we good at "predicting how a particular lesson or explanation or piece of feedback changes what a student understands and remembers later." Ultimately, I think, the article is a pitch to philanthropy: "Much of what we need, it turns out, is shared scientific infrastructure that no individual lab or company has much incentive to build today." 

This post summarizes a webinar organized in response to two open books, one on distributing (and distributed) knowledge (212 page PDF), and the other on the politics of open infrastructures (388 page PDF) (which include the challenging argument that the internet infrastructure has never been open). Both take an explicitly international perspective, and the webinar reflects this, looking at conditions of access in war-torn regions like Ukraine and systematically underserved regions in Africa. The different cases return to a common theme: "openness is not enough" (which is fair enough, but I hearken to my own principle that nothing is ever 'enough'). That said, as noted here, "the Distributed University invites us to see higher education not as a collection of autonomous campuses, but as a network of interconnected educational, research, and infrastructural capacities in which equity and openness become foundational principles." The recording is available to view online, and the slides are available to download.

This article (16 page PDF) is a response to a 2025 OECD report (26 page PDF) on "how to strengthen the role of research and initial teacher education institutions in knowledge mobilisation." Gert Biesta argues "the report assumes that the research is available but that for some reasons it doesn't reach policy and practice." There's an assumption that there is a sort of 'knowledge' - singular - around which there is evidence and consensus, and which practitioners should simply "make use of". But this represents a misunderstanding of 'practrice'. "A practice rather is collective activity held together, so we might say, by its purpose. And this purpose is not external to the practice." Different cultures, regions and industries have difference conceptions of this purpose, and you can't simply 'apply research' to them. Biesta describes this as "education's double history," corresponding to poiesis and praxis, but if we look at it, there are multitudes of histories. All in all, "what teachers do is never 'just' an instrument for bringing about certain 'effects.'"

In this article, "the authors propose the creation of integrated and open platforms that position OER as global public goods, grounded in open principles and standards." The argument is that developing these platforms could foster "inclusive, transparent and collaborative development, particularly in AI, on a global scale." Specifically, there is "an urgent need for open, transparent and publicly governed alternatives to commercial AI platforms" that would conform with UNESCO's ROAM universality indicators of "Rights, Openness, Accessibility, [and] Multi-stakeholder participation." I think there's a good argument for this, though I begin to resist when the authors go further, suggesting "the ethical deployment of AI requires adherence to ethical design principles, ensuring fairness, equity, transparency and protection of learners' privacy." I think they mean this as a premise, but in my view it's weaker than the argument it's intended to support. There is, as a matter of empirical fact, no consensus on ethical codes, and great ethical diversity globally. But there is also independently a need for social and political institutions that govern wide-reaching and consequential technologies like AI.