Victoria Hedlund

@GingaGill @hopkinsmmi @educationgovuk Thanks for the kind words. I think it’s particularly timely now we have #Moltbook

@hopkinsmmi @TeacherGenAI @educationgovuk Absolutely fascinating & concerning. 😬

Really interesting piece by @TeacherGenAI on personalised tutoring in light of the @educationgovuk announcement. #media-d42c395c-d6c3-4caa-b567-3012956d1625" target="_blank" rel="nofollow noopener">victoriahedlund.substack.com/p/sexy-ai-the-…

Expressions of interest now open for the GenAI in Teacher Education Summit 2025 @ University of Roehampton & online on 30th October. Register interest here: forms.gle/GQPTFd4fpoeVco… Get the word around!!

@DonaldClark @dylanwiliam Choice of pedagogy or modalities etc.

@DonaldClark @dylanwiliam Yes - I guess that involves the tech companies understanding good pedagogy. Although I do think if the user is given choice (or the illusion of agency) this could be a winning combo.

A review of studies of professional development in math and science finds a 1 sd increase in teacher knowledge is associated with a non-significant increase of 0.08sd in student achievement, though a 1 sd increase in instruction quality has 3x the effect: bit.ly/4kDfyHg

@dylanwiliam @DonaldClark @AlKingsley_Edu

AI made your lesson plan. Now get critical. Try Lesson Inspector Mini GPT to reflect, review, and improve your AI-generated lessons. Upload or generate a plan, analyse it with Teacher’s Standards or otherwise, and spot bias before it enters the classroom. bit.ly/45JSdyK

@dylanwiliam Hurrah!

Just realized that the text of my second inaugural lecture (at the Institute of Education, 2007)—Assessment for Learning: Why, what and how—is no longer being sold by UCL Press, so the copyright reverts to me, and I can make it open access, so here it is: bit.ly/AfLWWH.

@dylanwiliam @GenAIToolkit @DonaldClark Thank you! 🤩

@TeacherGenAI @GenAIToolkit @DonaldClark Certainly. Looks very useful. Thanks Victoria.

I’ve created ‘100 Ways of Using GenAI in Teacher Education’ - a guide of 100 ready made ideas to try from lesson planning, assessment and SEND to critical thinking and reflective practice. Let me know how you get on! gamma.app/docs/100-Ways-…

@emollick This is one of the reasons I think academia as we know it (haha) is becoming redundant. It’s too slow.

As someone involved in academic research on AI, it is notable to me that most of the key experiments showing the impressive abilities of AI on work, medicine, psychology, and so many other fields were done on GPT-4… a model that is now so obsolete that it is gone from ChatGPT.

Join our TeachEd AI Network if you’re in teacher education or ITE. We’ve just reached 100 members and had our second meeting today, where we looked at what’s working, what’s not, and what’s next. It’s light touch with no commitment! linkedin.com/groups/10065742

@airesearchtools Ah. This explains a lot! Shame, because 50 a week goes very very quickly….ive started reducing the iterations and seeing what I can condense into one prompt!!!

My o3 is already telling me to reduce the pace. This is why

Check out my latest article: Attention as a Commodity: Neurofeedback and AI in the Changing Landscape of Education linkedin.com/pulse/attentio… via @LinkedIn

This AI gender bias is too bad for words. Thanks ChatGPT

#SaveTheDate 📆 20 March, 5:30-7:00pm Dr Velislava Hillman: Educational Futures: Governing, Datafication and EdTech.    Following Veli's book presentation, there will be a panel discussion with Victoria Hedlund and Dr Sipho Morrison. eventbrite.co.uk/e/dr-velislava…

Group for GenAI in ITT/ITE on LinkedIn. Join if you’re interested: linkedin.com/groups/10065742

@dylanwiliam @justinskycak A normal distribution of attainment due to a normal distribution of ‘rate of learning’ which is due to a normal distribution of working memory/cognitive differences?

@justinskycak In the 1960s, John Carroll estimated that the fastest learning student in a mixed-ability class would learn at five times the rate of the slowest learning student, and my own work in the UK suggests a similar ratio. The range of achievement we see is therefore hardly surprising.

Speed of skill acquisition matters because time is limited. Once we enter a range where developing further expertise in a domain becomes overwhelmingly arduous, we tend to switch to other things that we (often, correctly) feel are a better use of our limited time.

@OpenAINewsroom @OpenAI There’s no gender equity here for scientific explanations

Today we’re partnering with the U.S. national labs to explore how AI can accelerate scientific discovery. Over 1,000 researchers will test frontier models, including OpenAI’s o3-mini, and help shape the future of AI in science. openai.com/global-affairs…