🎬 Lights, Camera, Action-Figure

📚 Knowledge builders

• Sora → This was Open AI’s video generation model, but it has expanded to be more a of media depository and creation part of ChatGPT. When you get through the pitiful sage verification check, you will be able to see images and videos that people have created and create your own images and videos too. Now some of these could be great stimuli for creative writing, but on the other hand, a quick glance through them just reaffirms how the boundary between historical reality could become blurred for our young people as demonstrated with this JFK selfie below.

• AI: The Double-Edged Sword? → As with all fancy tech demos we can quickly get caught up in the magic and not stop to think about consequences. AI is no exception. To save your inboxes from a near three thousand word essay, I have written a piece here that I hope can start a conversation around AI and its use in education.

🤖 Industry updates

• Improved ChatGPT Memory → Paid ChatGPT accounts will soon be able to take advantage of a new memory feature. When the update comes live, ChatGPT will remember details across all your conversations; not just the current one. If like me, you tend not to start a new conversation for a new topic, you will be able to ask ChatGPT what conversations you have had this week and it should let you know. While it is only available for paid accounts now, it is like that it will come down to free accounts sooner rather than later.

• How University Students Use AI → Anthropic have released some interesting research on AI use at uni. They plot the main uses into the matrix below, which I think is a very useful way to conceptualise AI use. Directive use asks the AI to produce something, while collaborative use supports the user in the production of the product. Directive involves very little thinking by the human, but collaborative use is more likely to include human thought throughout the process. In their research, 47% of use was focused on direct inputs.

  

  The Anthropic taxonomy for student-AI conversations.

✨ Fresh prompts

1. Action Figure → The latest AI trend seems to be creating AI action figures. It looks fun, but I would certainly caution the use of doing this without a solid reason for doing so. Image generation is a very power-hungry activity, so use sparingly. The prompt below should provide a useful start to getting what you want. Just remember to adapt the information in squared brackets and upload a few photos of yourself. Do not forget that once you upload those photos they could become part of the training data.

Create a full-body action figure of [insert name or persona], posed and displayed inside a transparent, original-style blister pack as if found on a toy shop shelf. The packaging should include retro or modern action-figure aesthetics [you may specify], complete with a bold nameplate, accessories list, and any taglines or character stats typical of collectible toys.

The figure should be holding or surrounded by 2–3 of their signature items or personal favourites [Describe what these items are in detail]

Ensure the accessories are arranged attractively within the blister pack and that the background insert complements the character’s theme. The final image should be vibrant, high-resolution, and ready for sharing or saving.