Stephen Downes

This short article references a 2022 paper that studies how the order of questions in a test impacts how well the test-takers do. "A perfectly logical test-taker should do equally well no matter the order of questions," writes the author, "but research shows that humans are influenced by the order." In particular, "students answered more... questions (correctly) when they began with easy questions compared to when they started with difficult ones." The relevant question to ask here is this: if the result depends so much on the order of the questions, what is it exactly that the test is measuring?

I often ask what it is we're doing when we're teaching and learning, that is, what counts as success? This paper offers some insight into the sort of question I have in mind. To 'learn' something is more than just to come to know that something is true. That's just memorization. No, as David Bourget would argue here, it's to grasp what is being taught. "It is one thing to believe something, and it is another to grasp it. For example, everyone knows that life is short, but most of us arguably do not fully grasp this fact." But what is it to 'grasp'? He argues, "we grasp to the extent that our thoughts are grounded in experience, whether occurrent or non-occurrent... , what we experience matters to how we reason because that is how we are wired: consciousness isn't a late addition to our minds; it is the
most central, causally potent form of mental activity." I think this argument works, overall. Image: Ding.

This paper (21 page PDF) is a review of papers proposing AI competence frameworks for teachers (CFT) and organizes them into five categories: integrating AI competencies in existing CFT; modelling new AI knowledge domains; process-driven; AI systems-driven; and competence level-driven. True to their field, the authors stress the need to theorize. "The empirical and design-based nature of AIED requires a solid theoretical foundation. The adoption of theoretical frameworks serves as a unifying force, fostering shared concepts and terminology among researchers and designers." But it's not clear there is (or is going to be) agreement on just what to theorize. This depends on what we want teachers to do, and as the authors note, "it is important to understand what kind of problems AI teachers' competencies will solve and what kind of solutions AI teachers' competencies convey."

This article offers three examples of the use of AI in a learning concext (it's a bit too early to call them 'success stories' since we don't really have a story of what counts as 'success' in this context). In one example, students are provided "with customisable prompts that they can cut-n-paste from a shared GoogleDoc to save them typing the prompt in." In the second, a 'learning lead' "used Microsoft Copilot to create imaginative discussion prompts for students." In the third, a senior lecturer aims "to normalise the use of GenAI" by putting it in the handbook and offering training sessions.

Since one assumes the 'President & Fellows of Harvard College' know better, we must conclude that their publication of articles like this is deliberately deceptive. Here's the set-up: Frederick Hess imagines himself in a conversation with an Uber driver about the plan to forgive billions in student debt. "I pay for this car," she (the Uber driver) said. "I paid for community college. These people borrowed money to go to college. They promised to pay it back. Why am I supposed to pay their bills, too?" Well, first, because living in a society isn't a quid pro quo. Second, at tax time an Uber driver is declaring business expenses like car payments, which offset the cost. Third, it is recognition that without addressing tuition fees college is something that could never be accessed by the working class, as used to be the case. And finally, neither students nor Uber drivers should be paying more in taxes than they get back: in a just society, those with the means (including Harvard and its $50 billion endowment) who benefit the most from society should be paying the freight.

"I take the position that a large portion of contemporary academic work is an appalling waste of human intelligence that cannot be justified under any mainstream normative ethics," writes Nathan Robinson. He's writing about philosophy but he could be writing about a lot of academia generally (not all, though, and that's an important distinction). I have asked this question of myself, and knew I'd never be able to just teach philosophy at an institution. It's more important than that. We discussed this morning how we achieve a 'work-life balance' and I replied that I don't have one: my life is my work, and my work is dedicated to making the lives of people better. Singer resonates with me: "we all know it would be wrong to watch a child drown in front of you and do nothing, so why is it okay to let children die far away from you and do nothing?" (P.S. I feel very self-conscious saying stuff like this so mostly I don't, but it's there, behind all my work.)