Does a new theory of dopamine replace the classic model?
The Transmitter: Neuroscience News and Perspectives,
2024/09/13
Though the field of education has its own versions of 'learning theory', neuroscientists are working their way toward an understanding of how learning actually happens. This article describes a new approach, called adjusted net contingency for causal relations (ANCCR) that better represents the relation between dopamine spikes and prior experiences than one base simply on prediction errors called temporal difference (TD). For a complex system such a neural network no simply theory will be completely accurate, though the formulation or this alternative may lead to new directions in research in both human learning and AI. Here's the study (11 page PDF) this article is based on. See also this recent article.
Web: [Direct Link] [This Post][Share]
Has ChatGPTo1 just stolen 'Critical Thinking' from humans?
Donald Clark,
Donald Clark Plan B,
2024/09/13
Donald Clark reports that "Critical thinking was one of the famous 21st century skills, that everyone thought AI could never solve. It just has." Seeing AI successfully solve math and reasoning problems is not surprising; once you've seen what it can do with computer code, similar progress would be expected in these other areas. But I'm going to withhold judgement for now because critical thinking is a bit broader than the formal methods involved in these disciplines. Judgements need to be made about meaning and intent. Now I'm not saying AI will never become an effective critical reasoner - in fact, I'm pretty sure it will, which is bad news for all the flim flam artists and propagandists out there. It will take time. But as Clark says, "Limitations being eliminated week by week."
Web: [Direct Link] [This Post][Share]
Strengths, Weaknesses, Opportunities, and Threats: A Comprehensive SWOT Analysis of AI and Human Expertise in Peer Review
Roohi Ghosh,
The Scholarly Kitchen,
2024/09/13
Roohi Ghosh provides us with a nice SWOT analysis presenting AI and human peer review side by side. It's a useful presentation because it doesn't simply idealize human reviewers and demonize AI, but rather, provides something like the same standard of assessment to each. The results are presented in an easy-to-use table (pictured).
Web: [Direct Link] [This Post][Share]
Most students are using AI for academics
Laura Ascione,
eCampus News,
2024/09/13
This post refers to a report from the Digital Education Council (spamware, as they want you to provide personal information before you can see it). According to the report, 86 percent of higher-ed students are already using AI, with 54 percent reporting using it weekly. Not surprisingly, "80 percent of students say their university's efforts to integrate AI tools have not met expectations, underscoring the need for institutions to understand how students prioritize AI tools and to integrate them in ways that align with students' expectations." I think it's safe to say the AI ship has sailed, which means educators will increasingly need to shift from 'resist' to 'adapt'.
Web: [Direct Link] [This Post][Share]
Ed Tech Startup Behind L.A. Schools’ Failed $6M AI Chatbot Files for Bankruptcy
Mark Keierleber,
The 74,
2024/09/13
This is an example of a failed AI company that promised to deliver to a school division. Pretty much everything that could go wrong actually went wrong, as there were questions about invoices and subcontracting, about salaries and expenses, about potential misuse of student data, and of course about the investors who pulled the plug before the company could collect on its big contract. This is less of an AI story, I think, than it is a business story, but it's not hard to predict that we'll see a lot more of this in the near future.
Web: [Direct Link] [This Post][Share]
Things I was wrong about pt 4
Martin Weller,
The Ed Techie,
2024/09/13
The latest thing Martin Weller is admitting he was wrong about is artificial intelligence. But for good reasons: "I was," he writes, "largely dismissive of it, partly because I was grounded in symbolic AI (expert systems and the like), and had not really monitored the rise of large, language models and generative AI." But of course, "in 2024, we can't deny it's a thing." Still, he writes, "I think the impact of AI is wildly exaggerated and we're probably heading for a bubble burst for all those companies investing heavily in it."
Web: [Direct Link] [This Post][Share]
The Friday Fifteen: September 13
HESA,
2024/09/13
HESA reveals a new format today, the Friday Fifteen. I like the idea of a weekly listing of small posts linking to news items. The sources in this first edition are strictly mainstream news: HEQCO, OECD, Pie News, etc. which suggests to me they're maybe using a clipping service of even something like Perplexity. Some items of note: a story about a 'private' university in North Korea, Russian tuition fees, and that old paean about universities (in the US) costing less than you think, because of discounts, via the Hill.
Web: [Direct Link] [This Post][Share]
The growing estrangement between universities and society — University Affairs
Paul Wells,
University Affairs,
2024/09/13
I have virtually no opinions in common with Paul Wells, but I have addressed the need for universities to become more relevant to communities, and this post seems to address that. It reads though more like a thinly veiled warning that universities are going to have to align politically. For example, "The easiest way to 'force' universities 'to do better' is to 'review the eligibility requirements for the receipt of federal research funds to ensure strong university governance'" which would more directly address, say, protest encampments. To me, this sort of argument underlines once again the need for universities to become essential in the lives of the people in the community, because only the community can support the university when its funding is being pressured.
Web: [Direct Link] [This Post][Share]
There are many ways to read OLDaily; pick whatever works best for you:
This newsletter is sent only at the request of subscribers. If you would like to unsubscribe, Click here.
Know a friend who might enjoy this newsletter? Feel free to forward OLDaily to your colleagues. If you received this issue from a friend and would like a free subscription of your own, you can join our mailing list. Click here to subscribe.
Copyright 2024 Stephen Downes Contact: stephen@downes.ca
This work is licensed under a Creative Commons License.