Stephen Downes

When a paper beings with a question like "Is AI always good for education?" I always worry. Nothing this complex is 'always' anything. And it makes me wonder why they're asked the question this way. Anyhow, the authors address this question from two perspectives: the potential for AI to improve curricula, and the potential for AI to improve access. The latter includes the possibility of personalization, automation and cost reduction. The authors argue in their very long conclusion, "put simply, the jury remains out on whether AI can meaningfully enhance higher education quality." Meanwhile, "In terms of widening access, AI is far from inexpensive." In a series of "takeaways for policymakers and institutional leaders" they offer a range of criticisms of AI, though I think more research and deeper consideration would have led to more plausible reflections.

Is this the one? "Keytrace lets you cryptographically link your atproto/Bluesky account to external accounts across the web." It's based on the distributed ID specification (DID) and you can use Keytrace to publish yours (if you have a Bluesky account, you have a DID, but there are other ways of getting one as well). I'm not posting a DID for myself just yet (Ben Werdmuller has done his) because I'm, working on a plan to generate my own DID in my own software.

I enjoyed this light treatment of subjectivism but as a refutation of the idea it fails utterly. The author traces its origin to the horrors of World War I, where people could quite rightly be justified in rejecting the idea of 'absolute truth' (one sees a similar response to the horrors of the Thirty Years War). But Arthur Krystal's association of subjectivism in philosophers with their emotional responses fails to justify the distinction between a trauma-free world of facts and trauma-fill world of emotions. You can't just say "the findings or conclusions of scholars and scientists existed independently of those who formulated them and those who interpreted them" and leave it hanging. Even if there is a 'reality' independent of our perception of it, there is an infinity of ways to understand that reality, and none of them has any a priori claim to being 'true'.

I think this article is mostly marketing for the National Academy for AI Instruction, "a five-year, $23 million partnership between the American Federation of Teachers and three of the largest AI developers - Anthropic, Microsoft, and OpenAI," but I also think it's an initiative worth noting. It's interesting to look at the article's use of language, for example, "We're in this race for teachers to get this knowledge" of more meaningful use of AI, said Randi Weingarten, the AFT president. "This will become the most disruptive technology in our time." And also, "A lot of teachers are doing this work at home, just wracking their brains." And also, "these tools, if gotten into the wrong hands, can be very dangerous for our students, for our profession, and for our jobs." And also, "there's still a lot of fear in the absence of federal guardrails on privacy, on safety, on disinformation, on academic freedom." This article is heavy heavy on the emotional language.

So what about LLMs training on free and open content? "I don't believe the answer is rejection," writes Hong Minhee. "I believe it's reclamation." The core question here, says Minhee, is that of who owns the models? "Who benefits from the commons that trained them? If millions of F/OSS developers contributed their code to the public domain, should the resulting models be proprietary?" OK, I take the point. But here's the thing. Why do we assume we're 'one and done' with models trained on open content. The content isn't going anywhere. Anyone can train models on this content (unless, of course, we lock it down, which would simply preserve the status of the existing proprietary models). We act today as though there could only even be one Anthropic Code or ChatGPT. But that's absurd. There are already many many properly open source models. We will be drowning in them! Open content ultimately means genuinely open AI - if we allow it to. Via Ben Werdmuller.

I have no idea whether this idea will take off, but I'm onboard with it. The idea is that people (real human people) who create their own websites can add a file called 'human.json' where they attest that their website is created by a human and vouch for other websites they know are created by humans. It should remind you of the old 'web of trust' and I now dub this the 'web of humans'. Using it is a three-step process: create a human.json file on your website (here's mine); put a link rel="human-json" into the header of your web page pointing to the human.json file; and then install the browser extensions for Firefox and Chrome to identify sites maintained by humans. Here are the full instructions. See also Alan Levine.