1. It was National Poetry Day here in the UK today. There’s a good website with a wide range of classroom activities (and much more besides) https://forwardartsfoundation.org/national-poetry-day/poems-and-resources/

2. High Quality Research Rarely Informs Classroom Practice. Why? is a piece for Education Week by Thomas S. Dee from Stanford Universityhttps://www.edweek.org/leadership/opinion-high-quality-research-rarely-informs-classroom-practice-why/2025/09 He’s talking about education in the USA but the scenario he describes, the disjunct between research and practice, is pretty much global.

When a firm and practice-relevant research consensus actually does exist, it frequently fails to influence education policy and practice. The ongoing contentiousness around how to teach reading provides a key illustration of this dysfunctional dynamic. In 2000, the National Reading Panel articulated an evidence-based consensus on reading instruction that, among other things, underscored the importance of promoting phonemic awareness and phonics among early readers as they learn to identify unfamiliar words. However, an EdWeek Research Center survey conducted nearly 20 years later found that a large majority of K-2 teachers—75 percent—instead encourage early readers to identify unfamiliar words using various contextual clues.

3. Here’s an open-access piece from Taylor & Francis on the use of AI in higher education, How does artificial intelligence compare to human feedback? A meta-analysis of performance, feedback perception, and learning dispositions by Rogers Kaliisa, Kamila Misiejuk, Sonsoles López-Pernas & Mohammed Saqr https://www.tandfonline.com/doi/full/10.1080/01443410.2025.2553639

Feedback as a tool to support learning has received significant attention in education. Hattie and Timperley (2007) define feedback as information an agent (e.g. teacher, peer, book, parent) provides regarding one’s performance or understanding. This information aims to bridge the gap between what is understood and what is aimed to be understood, guiding students towards achieving specific learning goals. Studies have shown that delivering feedback appropriately and promptly can improve students’ learning experiences and outcomes (Hattie & Timperley, ibid). However, with increasing enrolments in online and face-to-face learning environments, providing timely and appropriate feedback to large cohorts of students becomes difficult, if not impossible, for teachers or peers. Where teachers and students use educational technologies, automated and artificial intelligence (AI)-assisted feedback systems powered by advanced techniques offer the potential to provide timely, personalised, and data-driven feedback to students, allowing for timely interventions and corrections. Such real-time responsiveness can enhance the learning experience, as students are provided with actionable assessments that can be immediately incorporated into their study strategies (González-Calatayud et al., 2021). PDF below.

4. Two recent posts by Ethan Mollick on his blog, One Useful Thing

1. Real AI Agents and Real Work https://www.oneusefulthing.org/p/real-ai-agents-and-real-work

AIs have quietly crossed a threshold: they can now perform real, economically relevant work. Last week, OpenAI released a new test of AI ability, but this one differs from the usual benchmarks built around math or trivia. For this test, OpenAI gathered experts with an average of 14 years of experience in industries ranging from finance to law to retail and had them design realistic tasks that would take human experts an average of four to seven hours to complete (you can see all the tasks here). OpenAI then had both AI and other experts do the tasks themselves. A third group of experts graded the results, not knowing which answers came from the AI and which from the human, a process which took about an hour per question.

2. On Working with Wizards: verifying magic on the jagged frontier https://www.oneusefulthing.org/p/on-working-with-wizards

In my book, ‘Co-Intelligence’, I outlined a way that people could work with AI, which was, rather unsurprisingly, as a co-intelligence. Teamed with a chatbot, humans could use AI as a sort of intern or co-worker, correcting its errors, checking its work, co-developing ideas, and guiding it in the right direction. Over the past few weeks, I have come to believe that co-intelligence is still important but that the nature of AI is starting to point in a different direction. We’re moving from partners to audience, from collaboration to conjuring.

5. And, finally and in the interests of longevity, more yogurt is what we all need! https://theconversation.com/what-the-gut-microbiome-of-the-worlds-oldest-person-can-tell-us-about-ageing-266161