OLDaily

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

New Theory Suggests Chatbots Can Understand Text
Anil Ananthaswamy, Quanta Magazine, 2024/01/29

---

It has become a cliché to state that generative AI are "stochastic parrots." While this may be true of some systems, this article suggests that there is evidence that  the biggest large language models seem to learn enough skills to understand the words they're processing. "As these models get bigger and are trained on more data, they improve on individual language-related abilities and also develop new ones by combining skills in a manner that hints at understanding — combinations that were unlikely to exist in the training data." It does beg the question of what we mean by 'understanding', but when an AI can generalize beyond the training data and demonstrate mastery of skills such as self-serving bias, metaphor, statistical syllogism and common-knowledge physics, it gets harder and harder to say it's just parroting a response. Via Donald Clark.

Web: [Direct Link] [This Post]

What’s the point of school? (*according to 600 kids)
Ewan McIntosh, Medium, 2024/01/29

---

According to this report, "The top three responses for both educators and learners were: to get a job;  for the enjoyment of learning, and getting a 'good education'; and to develop friendships." 600 learners and 80 educators were sampled. The writing could be a lot clearer, and a lot less wordy, both in the post and the full report (28 page PDF) (for example, despite the ordering, it appears that 'develop friendships' was the top factor, though it's euphemistically described as "a notable purpose of Scottish education" in the report. These are contrasted with 'the four capacities', which the report refers to a lot but never defines (there's a single blurry icon listing them, but here's a much better description): successful learners, confident individuals, responsible citizens, effective contributors. The most important (and therefore buried) point was this: asked where those skills were best developed, students pointed most to extracurricular activities, such as clubs and sports.

 

Web: [Direct Link] [This Post]

Backgrounder - Canada's Big Banks
2024/01/29

---

According to this post Canada's 'Big 6' banks earned $58 billion in 2023 and $61 billion in 2022. It didn't cite sources so I dug around and was able to confirm the $61 billion figure with this report from KPMG. Why is this relevant? Last year, the average university tuition in Canada was $6,834, according to Statistics Canada. There were 1,174,000 university students in Canada. That's a lot, but the bank profits would have paid full tuition for 8,487,000 students. We could give every student in Canada free tuition while barely denting bank profits. Is this reasonable? Well, Canadian taxpayers did bail out the banks to the tune of $114 billion in the 2008 recession. Perhaps the banks could consider giving back to the system that sustains them. If we are to achieve the benefits of technology, including educational technology, we have to address the distribution of wealth in society. Otherwise, after all this work, we will have generated nothing more than a crisis.

Web: [Direct Link] [This Post]

After AI beat them, professional go players got better and more creative
Henrik Karlsson, Escaping Flatland, 2024/01/29

---

Can AI teach? Well, it depends on what we mean by 'teach'. For some, teaching is an explicit act of instruction. But I think that teaching is, for the most part, the practice of modelling and demonstrating superior skill. Case in point: "For many decades, it seemed professional Go players had reached a hard limit on how well it is possible to play. Then AI beat them." After this, the level of play in Go improved across the board. "After a few years, the weakest professional players were better than the strongest players before AI. The strongest players pushed beyond what had been thought possible." As Henrik Karlsson summarizes, "Something is considered impossible. Then somebody does it. Soon it is standard. This is a common pattern." But we need more than the mere example, we need to know the thinking process: the learning "coincides with the release of Leela Zero, an open source Go engine. Being open source Leela Zero allowed Go players to build tools, like Lizzie, that show the AI's reasoning when picking moves." This and the IHE link via Matthew Tower.

Web: [Direct Link] [This Post]

Arizona State announces a plan to give up on education
John Warner, Inside Higher Ed | Higher Education News, Events and Jobs, 2024/01/29

---

The title above is the original title of the article (still present as I write in the article's metadata). The display title has since been changed to "ChatGPT Can't Teach Writing". The context is an agreement made between Arizona State University (ASU) and OpenAI to make ChatGPT-4 available free to "approved students". John Warner shows admirable restraint in his response: "that sound you heard was thousands of souls of first-year writing instructors being snuffed, as though the Empire had fired a bolt from the Death Star." Why there should be thousands of first-year writing instructions at all, given that students have presumably completed secondary education, was not examined. Warner argues that writing is a uniquely human skill that chatGPT cannot master, let alone teach," because there is no genuine intent at meaning or communication behind the generation of that syntax." I would say the jury is still out on that. In my own experiences with chatGPT I would say it writes quite well. Wayback version, in case you are forced to sign in.

Web: [Direct Link] [This Post]

Call on the education sector to stop posting children's faces online
Defend Digital Me, 2024/01/29

---

I am sympathetic with the objectives of this group. Children are not in a position to decide for themselves whether their images appear online. And there's no doubt these images are abused. I have to ask, though, where these same objections were when children were exploited as actors, placed in advertisements and appeared in news reports (Kim Phuc Phan Thi comes to mind). This suggests to me that the problem isn't really the exploitation of children's images, but a way to shift responsibility from the AI companies manipulating and misusing them to the schools benignly sharing their images online. Via Doug Levin.

Web: [Direct Link] [This Post]

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