OLDaily

What to Do When AI Fails
Andrew Burt, Patrick Hall, O'Reilly, 2020/05/19

The message seems clear enough: "To date, at least 1,200 reports of AI incidents have been recorded in various public and research databases. That means that now is the time to start planning for AI incident response, or how organizations react when things go wrong with their AI systems." But it begs an interesting question: "What is an incident when it comes to an AI system?" After all, when an AI  created Tay, a racist Twitter user, that was the AI performing as designed. This article looks at various types of AI failures, how they go wrong, what the consequences might be, and how to plan an analytics failure response.

Web: [Direct Link] [This Post]

Never Go Back to the Office
Juliette Kayyem, The Atlantic, 2020/05/19

I know, a lot of people can't work from home. But a lot of people - like me, for example - can. And we're beginning to see the benefit of leaving those who can work from home where they are. Even if we did go back, the offices would be very different - literally sterile, socially distant, hard to use. But many of us (including me) prefer it. Not surprisingly, " A Gallup poll last month showed that most people who are working from home want to keep doing so after the crisis abates." So - let's assume this is what happens. No matter what, it's arguable (as suggested here) that corporate education will never go back to the classroom. Even if we're in the office, we don't want to stuff ourselves into a roomful of people coughing and sneezing their personal collection of viruses for the rest of us.

Web: [Direct Link] [This Post]

OER for Workforce Development
Meg Lloyd, Campus Technology, 2020/05/19

This article links us to SkillsCommons, "an enormous repository of free and open source workforce development training and education materials." I tried a few searches ('philosophy', 'extrusion') and was pleasantly surprised by the results. You can also browse by industry, credentials and material type. Look at "Plastics and Rubber Products Manufacturing", say. That's where some weaknesses become apparent - the articulation agreement with Mississippi colleges is simply not useful, nor are the blended or hybrid courses (since you have to be there in person). Or, in a single resource, you might find a long list of slide decks. Similarly, a course preview isn't useful as an OER. I know, it's hard to get this right, and there are many competing interests.

Web: [Direct Link] [This Post]

Radical Solutions and Open Science
Daniel Burgos, Springer, 2020/05/19

A better title for this book would perhaps have been 'Open Education and Open Science', and in fact, it mostly focuses on the education part. No matter: the subject of this open access book (199 page PDF) matches almost perfectly my own range of interests. Topics covered include open access, open science, ethics in analytics, OERs, open pedagogy, MOOCs, and related issues. The papers don't break a lot of new ground and it's a quick read, but the coverage is good and this would be a useful text for people new to the field. Highlights for me included the distinction drawn between the ethics of operations research (OR) and academic research by Dai Griffiths as well as the question of whether OERs are contributing to the private appropriation of the educational commons raised by Tel Amiel and colleagues.

Web: [Direct Link] [This Post]

Life skills, technology, teacher evaluation and the World Bank
Philip Kerr, Adaptive Learning in ELT, 2020/05/19

This is a very good examination of the role the World Bank plays in language learning, and by implication, in learning in general, especially in the global south. The key questions asked here are (1) what counts as success for the World Bank's education programs, and (2) success for whom? It is arguable (and argued here) that the World Bank is as interested in lowering wages as it is in promoting economic development. "The teaching of ‘life skills’, the promotion of data-capturing digital technologies and the push to evaluate teachers’ performance are, then, all closely linked to the agenda of the World Bank:

• cost accounting and quantification (since returns on investment must be carefully measured)
• competition and market incentives (since it is believed that the ‘invisible hand’ of the market leads to the greatest benefits)
• the private sector in education and a rolling back of the role of the state (since it is believed that private ownership improves efficiency)."

The principle are, of course, quite familiar to most observers of education policvy generally. These same observers should be aware of the harm they have produced and how poorly prepared they leave us should, say, a global pandemic completely upset the world economic order.

Web: [Direct Link] [This Post]

Educational Crises and Ed-Tech: A History
Audrey Watters, Hack Education, 2020/05/19

What I like about this post is that it recounts a history of the use of technology to replace traditional schools during times of crisis and disease, thus proving that our times are not exactly "unprecedented". Unfortunately the history is entirely US-based, and essentially (to my reading) boils down to the well-worn argument that "previous uses of tech to replace schools didn't work, so we shouldn't expect anything different this time." And maybe not, in the U.S., where many other factors are at play in addition to disease. But what about countries with a solid history of on-air schooling, of educational television, of distance learning institutions? Maybe something different. Maybe something better.  (Also on NEPC yesterday).

Web: [Direct Link] [This Post]