news (external)

There are two really important points in this short article. The first revolves around the need to fix online authentication. Passwords are awful, but everything else that's been tried is worse. We will eventually get there, but it will require companies to allow users to own their own logins. They won't do this - they're like the monkey with its paw stick in the peanut jar, too greedy to open their fist and let their prize go. The second reflects our frustration over the whole state of affairs. "What I choose to do is irrelevant, because I'm at the mercy of the people who create the sites I use. And their practices can be shockingly bad." We see the same pattern everywhere: we (as consumers) know what would be good for society, or the environment, or whatever, but we can't choose it because the option is just not available in the marketplace.

"these words are almost always used to defend traditionalist practices such as direct instruction and control-based interventions," writes Alexandra Mihai. By contrasts, getting to in-person conferences take a lot of planning and time. "It feels like these constraints make our experience poorer." You are also, I would suggest, seeing different people, specifically, those with means and motivation to travel. This shapes in a negative way what gets valued and what gets decided. "Today's newsletter is a call to try and bring that spirit back." We need to create conditions for genuine online interaction with or without in-person conferences. I mean - I get the appeal of in-person. Travel enbiggens the mind. But I also feel (as I have the last few years) the feeling of apartness one gets when the travel opportunities dry up.

I'm totally on board with what Alfie Kohn is saying here. The overview (quoted):

• some people take an extreme, reductionist view of what qualifies as data.
• what exactly is meant by "effective"?
• evidence of an "on-balance" effect for a given intervention doesn't allow us to claim that it's a sure bet for all kids.
• what's sold to us as "evidence based" can sometimes do more harm than good.
• the term evidence-based sometimes functions not as a meaningful modifier but just as a slogan
• these words are almost always used to defend traditionalist practices such as direct instruction and control-based interventions

Gain attention, create motivation, maximize engagement. That's the 'ultimate online learning blueprint' offered here.

This paper (22 page PDF) offers a systematic review "to provide a comprehensive perspective on the current state of research and practice, and to offer insights on how ChatGPT might be effectively integrated into K-12 education." Specifically, it presents a Strengths, Weaknesses, Opportunities, and Threats (SWOT) analysis of ChatGPT, as well as "recommendations concerning ChatGPT practices within the realm of K-12 education." The results are not even slightly surprising: "ChatGPT demonstrates a level of performance comparable to that of humans in a wide range of domains, establishing itself as a flexible tool for educational purposes. Nevertheless, the study also recognises some constraints of ChatGPT."

This is note directly an edtech story but it has obvious implications in the edtech space. As documented here, "Previously unreported documents suggest age assurance technology has not been successfully implemented anywhere in the world." Now I've gone through verification processes several times - once to get a Twitter blue check back when it mattered, once to open a blockchain account, for example, and another to register at a contact-less Sonder hotel. I never felt confortable with it and was quite aware of the potential for abuse. There's no third-party solution that can be trusted not to misue the data - and that applies especially to the financial sector (even though governments have started using banks to verify ID and many other services require a credit card check).

This is a discussion those involved in free and open content should be aware of and have an answer for. Here's the gist: Sue Smith writes, "I am begging ethical web enthusiasts to understand what an extreme privilege it is to spend time working on something without worrying about money." And Marco Rogers writes, "Too many people struggle with this binary. If you care about money at all, there's only a short leap in people's heads to the worst kind of exploitative behavior. And the only way to avoid that is to not care about money at all." The full comment threads for both are worth reading. My feeling: if you are unable to share your thoughts and feelings without getting paid, because (say) you're too poor, then you don't own your own voice. It belongs to whomever is paying. People should never be in such a state of want that they have no voice. The solution here isn't to pay for content; that just entrenches the inequality and privilege of the few. It's to address the wider issue of poverty, to what is today a privilege becomes a right tomorrow.

This study (18 page PDF) offers " information about the current state of emotion recognition technology and how its practical use is being carried out in educational settings." It does point out that "most of the research has been conducted from a theoretical perspective and none of them has been fully developed and implemented in the classroom." This is no surprise - the classroom is the last place I would expect to see educational technology deployed. Anyhow, it considers methods like facial recognition, speech analysis, and combinations of these and other cues. About 14% of the systems were at technology readiness level (TRL) 5,6 or 7 - "validation and demonstration in a relevant or operational environment" - and none were above that.

Here's the summary: "This paper (27 page PDF) presents a framework for grappling with openness across the AI stack. The paper surveys existing approaches to defining openness in AI models and systems, and then proposes a descriptive framework to understand how each component of the foundation model stack contributes to openness." I appreciated the comprehensive approach to 'openness in AI' found in the document; unlike (say) OSI, it doesn't endorce one 'received view' of openness, but rather looks at the many different ways AI can be open or not open.

This is a good article and I think people will like it, though I am less comforted. The authors argue they want to go beyond a technology-led approach to pedagogical reform to improve access and social justice in education because of the failure of technology alone to accomplish any of these things. I can certainly understand their frustration; academia has been remarkably impervious to any change that might redress historical inequities in education. The authors offer as an alternative "a virtual makerspace and a guided fantasy story that were applied in a project concerned with rewilding higher education pedagogy." I laud their intentions but I'm not seeing how this is going to create the change they are looking for. You can't reform the existing system by changing the dressing on the walls. But hey, maybe I'm wrong, and maybe you can convince academia to care.

People haven't really been talking about  'quantum ed tech' yeat and probably won't in the near future, but from what I've seem quantum computer technology, quantum algorithms are really weird, and as this article suggests, there will be a quantum internet some time in the future. The argument here is that while government has been funding basic research, it needs to fund industry in order to develop practical applications. Personally I think that's a bad idea - it's too early for that, and the companies will just take the money without producing any tangible result. And when the technology does reach the stage of practical application, companies will jump in with their own money, as we've see with AI. Let's first get the technology right, and maybe even think ahead of time about how it will be managed.

I guess I shouldn't be surprised by the alarm this feature is raising: "At a Build conference event on Monday, Microsoft revealed a new AI-powered feature called "Recall" for Copilot PCs that will allow Windows 11 users to search and retrieve their past activities on their PC. To make it work, Recall records everything users do on their PC, including activities in apps, communications in live meetings, and websites visited for research." This is the trade-off, though. To provide these great services, the AI will have to have access to your personal information. It can't just make stuff up! You can and should be able to turn it off. But then you have to do without the service. But that just cedes the advantage to people like me who will turn it on and have access to these far-reaching capacities. Privacy! Hah! I grew up in a small town, I know how to live without privacy.

This is a project from a long-time member of Tim Berners-Lee's SoLiD project, Sarven Capadisli. "dokieli is a clientside editor for decentralised article publishing, annotations, and social interactions." It has all the strengths and weaknesses of the SoLiD project itself: it looks great, has some really neat features, but the user interface isn't very friendly, the documentation is unhelpful, it has been years in development without deployment, and it can't actually be used by people for practical purposes. But do have a look - the web page is also a content editor, and there are some nifty features.

The website EdTechnica describes itself as "The Open Encyclopedia of Educational Technology" and is part of the EdTechBooks website. It contains about 30 articles on ed tech topics written by various authors. This article describes the project and asserts, "As an OER, EdTechnica extends the 5Rs of openness—retain, reuse, revise, remix, and redistribute by also attending to the 3Rs of social justice—recognition, representation, and redistribution." The articles are of acceptable quality, though over time I would expect to see more rigour and standardization.

According to this article, "Compared to traditional video, environmental stories told through metaverse technologies, including virtual reality and 360-degree video, can better motivate people to act on environmental threats." It's based on a research report (8 page PDF) by Daniel Pimentel and Sriram Kalyanaraman. All very well but it should go without saying that if VR can motivate environmental activism, it can also use misinformation to do the opposite.

To quote Matthew Tower, "A new type of homework assignment: write an essay that is better than ChatGPT's answer to the same prompt." According to the article, "Students in two courses at the University of Nevada, Reno, are going head-to-head with ChatGPT by answering the same prompts as the AI and aiming to get a higher grade." Tower finds this "super compelling" because "You have to 1) understand the assignment, 2) are effectively deterred from using ChatGPT because the ChatGPT answer is a given, and 3) have to think about what makes your response 'better' than the stock robot answer." I have to admit, it's creative, even though the assignment might boil down to finding a better prompt than the one the instructor used.

As reported by the Communications Workers of America (CWA), "Berea College is a work college and requires all students to work on campus." Working conditions, however, are terrible. "The schedule left me struggling, sleeping through my morning classes, and failing a class that I had to drop," said one student. "If you get stuck with a certain position and have no voice on the job, there's a correlation and a causal relationship with drop-outs." After several days, the Labor Relations Board hearing has finally started. According to the Chronicle (paywalled) the College calls it an "Existential Threat." If your existence depends on exploiting student labour, then maybe it should be questioned. More: Fox56, which points out the students don't pay tuition; Lex18, which reports "Our message is it is pro-Berea to be pro-union," said Andi Mellon, a Berea student. Berea hasn't charged tuition since 1892 and has an endowment "worth around $1.2 billion, and profits from the investments cover a large portion of what it costs to educate more than 1,600 students.

As Victor Mair reports, "the fake science sickness has infected some of our mainstream publishing  houses." It's easy to blame AI for this, but of course it's not just AI. It's the people abusing AI to try to tap in to the the nearly $30 billion academic publishing industry. And they've always been around. Mair argues that the law should get involved. "Accredited authorities should go on the offensive and work for the enactment of laws and penalties," he writes. "Make these crimes of sham scholarship cost." More effective, I think, would be to take the money out of the system. Colleges and Universities could publish in-house and make the papers open access. Then there's no way to make money out of the system, and the people who pay for the existing system - authors and institutions, who would now be publishing in-house - end up paying a lot less.

Remember when people told you "what's online is forever?" Well, according to this Pew report, "A quarter of all webpages that existed at one point between 2013 and 2023 are no longer accessible." I have no reason to doubt that, based on my own experience. "23% of news webpages contain at least one broken link, as do 21% of webpages from government sites.... 54% of Wikipedia pages contain at least one link in their "References" section that points to a page that no longer exists.... Nearly one-in-five tweets are no longer publicly visible on the site just months after being posted." Via Dogtrax.

Just a quick update from David Wiley on the Open Source Initiative (OSI) attempt to own, er, I mean, steward, the definition of 'open source AI'. Wiley notes "The definition is currently in its eighth draft, with the goal of finalizing the definition by October, 2024." Why am I so sceptical of the OSI initiative? Well, the first sentence contains the phrase "massive benefits accrue" and everything else is based on this, as though there would be no point to open AI if it didn't make money. Yet (ironically) there's no requirement in the four freedoms (use, study, modify, share) that these benefits be distributed equitably, or that they not deprive one of any existing freedom, or cause no harm in their application. Moreover, 'open source AI' is "made available", and not (say) developed openly, or developed by a community; it's as though it's some sort of creation that comes down from on high. Defining 'open' economically, and then cleaving it from social justice and any sort of community-based process, is the sort of 'open' we might find in Silicon Valley, but not really one we want in our own communities.