🏫 EducAItion, educAItion, educAItion

Have you asked how pupils use AI?

Author

Neil Almond
September 9th, 2025

👋 human,

I hope all is well in your school. I really enjoyed the replies to my previous question around safeguarding and AI; it was interesting to hear about the range of experiences people were getting around this sensitive topic. Here’s what’s new this week on all things AI:

📚 AI+education news

  • OpenAI for Greece > Greece will pilot ChatGPT Edu, a version of ChatGPT for academic institutions, in upper-secondary schools. The pilot, led by the Onassis Foundation, aims to build AI literacy and integrate AI responsibly into teaching.

    It still strikes me with concern, but perhaps no surprise, how quickly AI is being put into education at such a scale without any high-quality research into the medium and long-term impact.

  • EducAItion, educAItion, educAItion > The Social Market Foundation has released a report warning that if generative AI tools are used uncritically, there is a risk of disrupting or even circumventing the acquisition of new knowledge and the enhancement of existing knowledge. This could lead to poorer outcomes for learners of all ages. You can download the whole paper here.

🌍 Wider AI updates

During 50 hours of testing using accounts registered to children ages 13-17, researchers from ParentsTogether and Heat Initiative identified 669 sexual, manipulative, violent, and racist interactions between the child accounts and Character.ai chatbots. That's an average of one harmful interaction every five minutes.

Mickey Carrol - Sky News Reporter

  • Branch In New Chat > One of the issues with the Chatbot interface for the user is the linear structure that each server and return provides. This means responses or ideas that you want to follow up on quickly get lost. However, a new web feature for ChatGPT means that you will now be able to branch interesting responses into new chats to help organise chats. Just click the ellipsis at the bottom of a reply and click ‘Branch in new chat’.

🎯Prompt

  • No prompt this week. Just a framework that I hope to explore in more detail as this year goes on. To make sure you get the most out of your prompts, make them CRISPY.

    • Clear structures > Use specific, unambiguous language and syntax.

    • Role-based > Tell the AI who or what to think like.

    • Input-specific > Set precise parameters and constrains

    • Steps-based > Provide a logical sequence to tackle tasks.

    • Purpose-driven > Define a clear objective for the output

    • Yield-focused > Specify the desired output format.

Till next week.

Mr A 🦾

Reply

or to participate.