OLDaily

Welcome to Online Learning Daily, your best source for news and commentary about learning technology, new media, and related topics.

How I Use "AI"
Nicholas Carlini, Writing, 2024/08/16

---

Since the release of the first GPT models people have been critical of AI's ability to do anything practical and in many cases making the case that it's just another fad. I don't think it is, and neither does this author, and our reasoning is similar: "But the reason I think that the recent advances we've made aren't just hype is that, over the past year, I have spent at least a few hours every week interacting with various large language models, and have been consistently impressed by their ability to solve increasingly difficult tasks." Now I've spent a lot less time with the AI. But O am incredibly impressed not only with its ability to perform specific tasks but also in how it supports my own learning about those tasks. I'll write about that later (as a human, not with AI - I reserve AI for coding tasks); in the meantime, you have this article.

Web: [Direct Link] [This Post][Share]

Can astrologers use astrological charts to understand people's character and lives? Our new study put astrologers to the test
Spencer Greenberg, André Ferretti, Clearer Thinking, 2024/08/16

---

What I really like about this article isn't what it has to say about astrology but rather the way the authors are clear that this is an example of how we test a claim scientifically, by making the claim precise, choosing a specific measurement, and having clear criteria to describe what success would look like. We see some practical applications that tested not only for predictive accuracy but in comparisons based on the astrologers' own confidence, (self-)assessments of skill levels, and consistency across different practitioners.

Web: [Direct Link] [This Post][Share]

Balancing Online Trustworthiness and Anonymity with Personhood Credentials
Decentralized Identity Foundation - Blog, 2024/08/16

---

This proposal (63 page PDF) has the potential to be important. The authors describe a 'personhood' credential to help websites distinguish between humans (and perhaps their authorized AI agents) and bots. An attractive feature of the proposal is that it preserves anonymity; all the website knows is that you're a person, but not who. The credential depends on in-person verification (but UK take note: not a pub) and cryptography. Would it work? Probably. But would it be abused? Almost certainly. It's a tiny leap from verifying in person that you're a person to verifying in person that you're an adult. And from there to verifying that you're, say, a citizen. Or a CAA member. Or a political party member.

Web: [Direct Link] [This Post][Share]

The AI Sounding Board
Philippa Hardman, Dr Phil's Newsletter, Powered by DOMS AI, 2024/08/16

---

There's still a couple steps to go from this, but we have here a first indication of how we get to what I've called 'deontic analytics', that is, AI that is able to determine what we should or ought to do. Here's Philippa Hardman: "Imagine a world where, with the right inputs, AI can reliably take on the role of business leaders and target learners for infinite conversation about the problems we're trying to solve and the best-fit solutions. The answer is: yes, with some important caveats." Now of course, stakeholder feedback isn't the same as ethics. But the two are not that far apart. The caveats are, of course, what we might expect: lack of insight into psychometrics, demographic bias, and responses that trend toward the dull and middle path.

Web: [Direct Link] [This Post][Share]

Global Data Consortium: Artificial Intelligence’s Essential Role in the Future of Universities
George Siemens, Srecko Joksimovic, Damien Coyle, Walter Reilly, Shane Dawson, Rose Luckin, American Council on Education, 2024/08/16

---

This paper (15 page PDF) proposes a Global Data Consortium (GDC) "to democratize access to educational data and leverage the potential of AI to address disadvantages and enhance learning and teaching outcomes for all." According to the proposal, "Participating institutions will gain access to a secure data infrastructure, a community of experts, AI models and algorithms, and opportunities for collaborative research and professional development." The paper cites as examples MOOCdb, which "provided a standardized data schema for data extraction from EdX and Coursera courses," and the MOOC Replication Framework (MORF), "a platform-as-a-service consisting of data as well as experimental and inferential infrastructure." I'm sure I'm not the only one to look at this and see a proposal for universities to pool their data in order to sell it collectively at the highest rate possible. This is from last May but only showed up in my RSS reader today.

Web: [Direct Link] [This Post][Share]

AI 'Companions' are Patient, Funny, Upbeat - and Probably Rewiring Kids' Brains
Greg Toppo, The 74, 2024/08/16

---

Loaded commentary about the rising trend of chatbots for children (I've documented a few of these recently). Greg Toppo uses the 'one observer' device to editorialize while pretending to report news: "The proliferation of friendly, often sexually suggestive, chatbots is 'making it so easy to make a bad choice,' one observer says," he writes. To be clear - everything rewires kids' brains. The more exposure they get, whether to television, nature or chatbots, the more they will learn from that exposure. That's why simply banning chatbots won't help - if they put down the chatbot and start watching Fox News, they're no better off. What we want are chatbots that promote healthy attitudes and habits of mind. The best way to do that is to build them, rather than complaining about them. But hey, why do that when you can write a good scare story?

Web: [Direct Link] [This Post][Share]

E-Mail Done My Way, Part 0 - The journey
Jan Wildeboer, Jan Wildeboer's Blog, 2024/08/16

---

This is an absolutely delicious series of posts in which Jan Wildeboer describes how he sets up his own personal email servers using a minimalist configuration. As he notes: "This whole series is not a simple HOWTO. This series is about how I run my mail server. I go through the configuration, line by line and explain. It's not going to be a simple Copy/Paste to run your own mail server. You have been warned." Still: if you want to understand the sorts of things that email servers need to do - everything from assigning domains to filtering soam to sending email to modern server protocols like DMARC - then this series is for you.

Web: [Direct Link] [This Post][Share]

Personality-sensitive pedagogies: A study of small group interactive behaviours among 9- to 10-year-olds
Marcus Witt, Ben Knight, Tom Booth, British Educational Research Journal, 2024/08/16

---

The authors emphasize that this and that we should treat any conclusions with caution, but as someone always wondering what the learning-styles sceptics would say, I found this an interesting result. "Personality is a useful factor in examining group processes," suggest the authors, "and can lead to insights for classroom practitioners to increase efficacy and equality of participation." Moreover, "there is evidence to suggest that unequal participation in collaborative learning situations has an adverse effect on learning." At a minimum, work like this suggests that the relation between personality and learning outcomes is more complex than teach-then-test instructivists may suggest.

Web: [Direct Link] [This Post][Share]

We publish six to eight or so short posts every weekday linking to the best, most interesting and most important pieces of content in the field. Read more about what we cover. We also list papers and articles by Stephen Downes and his presentations from around the world.

There are many ways to read OLDaily; pick whatever works best for you:

• Read in your web browser on this web page, updated weekdays
• Receive a daily or weekly email newsletter by subscribing here
• Subscribe to the RSS feed using your favourite feed reader
• Follow @OLDaily@mastodon.social on Mastodon or elsewhere in the fediverse
• Follow @OLDaily on Twitter
• Integrate the open API using the JSON feed