Hi
Hope you had a wonderful weekend wherever you are - mine was spent chilling in the hills outside of Hanoi in Ba Vi National Park. Wonderful getaway weekend for me but AI is very much no days off - so, with that, a few headlines from the world of AI:
🤖 Claude Gets Physical: AI Takes Control of Your Computer
Anthropic have just dropped a sneak peek at the future of agentic AI - they’ve released a public Claude beta that can use computers the same way people do.
What does that mean in practice? It means AI looking at a screen, moving a mouse, clicking buttons, and typing text - while it’s a bit clunky as yet, this is absolute “we-live-in-the-future-madness” - see it in action here #gamechanger
Unsurprisingly this has blown up in the edtech x AI space: Donald Clark calls it "practical agency", Leon Furze's tested it extensively, and Ethan Mollick's gone deep on what it can do. Think "next ChatGPT moment" for education - and I'd encourage people to take this as seriously as it deserves. After all, there's essentially no such thing as a non-AI online assessment anymore. Any computer task? AI can now do it 😵💫
Oh and Anthropic dropped an absolutely fantastic analysis tool which lets you make interactive data visualisations with Claude artifacts - just because they can, I guess 😲🤩
🌍 Green AI Revolution: 95% Less Energy, 100% Performance
Game-changer alert: a new paper drops the mic with "Addition is All You Need" - they've potentially cracked the code on energy-efficient AI. By swapping complex math for simpler calculations, they're getting 95% energy savings with zero performance loss. 🌿
Why this matters? Students are using AI like crazy - 8000 Aussie students and 4000 global students show it's basically their new search engine. The catch? AI's current water usage is insane (at least 500ml per chat session?! and trending toward UK-level consumption by 2027) - sustainable AI is very much the way forward.
This breakthrough couldn’t come at a better time for education...
🎯 The AI Impact Navigator: Making Theory Work in Higher Ed
Australia's National AI Centre has developed an interesting framework that might actually help universities tackle AI implementation. The AI Impact Navigator moves beyond theoretical frameworks to provide practical steps for assessing and improving how we use AI in education. Built around a "Plan, Act, Adapt" cycle, it focuses on four key areas:
✅ Making AI work for the whole campus community
✅ Protecting student experience and rights
✅ Enhancing staff and system effectiveness
✅ Building trust through transparency
The Navigator helps universities move from vague AI initiatives to measurable outcomes - shifting from internal metrics to real, reportable results. It's designed to help build diverse teams, protect privacy, maintain academic integrity, and create responsible AI practices that work for everyone.
But is this really different from previous frameworks, or just a fresh coat of paint? Let's look at something more radical...
🔄 The Great Unlearning: Education's AI Revolution
The Great Unlearning hits hard: we need to stop stuffing heads with facts and start teaching adaptability. Carlo Iacono argues for a radical shift - "learn, unlearn, and relearn" is the new game. Out with static knowledge, in with dynamic thinking that can keep pace with AI.
Donald Clark* takes this further, targeting assessment head-on. Traditional tests? Dead on arrival. When AI can ace any online task (hey Claude! 👋), we need authentic, performance-based evaluation that makes AI an ally, not an enemy.
Perhaps it's time to stop playing cat-and-mouse with AI and start revolutionising how we teach and test 🤖🎓
*Twice in a single edition - that's a new record!
🎬 AI Creative Suite: Midjourney's Photoshop Killer & Runway's Actor Generator
Madness from the creative space with Midjourney now allowing you to upload your own images to their models and image editor - overview here and a video rundown here (some big calls in there vs Photoshop #shotsfired).
Runway have clearly had enough of KLING AI chewing on their heels and have released Act-One - a new way to generate expressive character performances inside Gen-3 Alpha using a single driving video and character image. What does this mean? Low-budget independent filmmaking revolution? ✅ International content localisation? ✅ Historical/biopic productions? ✅ Age-fluid storytelling? ✅ Exciting stuff - creative education will never be the same! 📺👀🍿
We're seeing a convergence of transformative AI developments in HE and beyond. From Claude's ability to control computers (challenging traditional assessment) to breakthroughs in energy efficiency (making AI more sustainable), the landscape is rapidly evolving. While frameworks like the AI Navigator offer practical guidance, the real challenge lies in fundamentally rethinking how we teach and assess students. Meanwhile, creative tools are opening new possibilities in educational content creation. Wild times for the world - and for HE in particular.