Hey {{ first_name | human }},
This week we are serving up some delicious updates and a slice of AI Literacy Assessments coming to PISA in 2029.
TL;DR: The 60 Second briefing
⚡️Fable’s back: Anthropic has restored access to Claude Fable 5 and Mythos 5 after US export controls were lifted.
🧪AI Tutor companies revealed: The DfE has announced the partners for its safe AI tutoring tools for Years 9–10 with potential national availability from 2027
🚨OpenAI slows GPT 5.6 rollout: OpenAI’s GPT-5.6 family, Sol, Terra and Luna is starting with a limited preview before wider release, after engagement with the US government.
📚 AI+education news
🧪 AI Tutor companies revealed: > What it is: The government is supporting the development of AI tutoring tools for disadvantaged pupils, with the aim of making safe, curriculum-aligned tools available to schools by the end of 2027. The DfE says these tools should complement high-quality teaching rather than replace it. Schools Week reports that eight suppliers were selected: Google, Pearson, Eedi, ElevenLabs, Flint, Learn Anything Education, Medly AI and Zero Gravity Tech.
Why this matters: AI tutoring is moving from private experimentation into public procurement. That means the important questions are no longer just “Does it work?” but “Who oversees it?”, “How is it aligned to the curriculum?”, and “What happens when pupils get stuck or disclose something concerning?”
Do this next: Ask any AI tutoring provider three questions:
1. What curriculum is it built around?
2. What can the teacher see?
3. What does the system do when it is unsure, wrong, or worried about a pupil?
⚡️ AI rollout fragmented > What it is: Teach First and Accenture have published a report on AI in schools. Their warning is that AI adoption is already happening, but often in fragmented or reactive ways. They argue that disjointed adoption could widen inequalities between schools and pupil.
Why this matters: The AI divide may not simply be about who has access to tools. It may be about which schools have the leadership capacity, staff confidence and shared routines to use AI well.
🧪 PISA launches MAIL assessment > What it is: PISA 2029 will include a new Media and Artificial Intelligence Literacy assessment. The OECD says this will look at whether pupils can engage critically, ethically and responsibly with digital content, media platforms and AI systems. Results are expected in December 2031.
Why this matters: It is clear from the CAS review and PISA that AI literacy is here to stay. The question becomes how it gets integrated into curriculum across schools and at what cost.
🌍 Wider AI updates
⚡️ Fable is back > What it is: Anthropic has restored access to Claude Fable 5 and Mythos 5, its latest models, after US export controls were lifted. Access had been suspended because the rules applied immediately and Anthropic said it could not reliably verify users’ nationality in real time. Fable 5 is now returning globally, while Mythos 5 access is being restored more cautiously for approved partners.
Why it matters: This is a sign that the most capable AI models are no longer being treated like ordinary software updates. Governments are beginning to see frontier models as strategically important infrastructure, especially where cybersecurity and national security are concerned.
🚨OpenAI slows GPT 5.6 rollout > What it is: OpenAI has started with a limited preview of GPT-5.6 Sol, Terra and Luna before broader release. The company says this followed engagement with the US government, and that access is initially limited to a small group of trusted partners. OpenAI also says GPT-5.6 has stronger coding, science and cybersecurity capabilities, alongside a more robust safety stack.
🎯AI concepts every teacher should know: 4. Transformer:
No, no. Not Optimus-Prime-style-transformer . It’s what the T in ChatGPT stands for (Generative Pre-trained Transformer).
So far we have looked at tokenisation, embeddings and attention. An embedding gives the model a starting sense of what a token could mean. Attention helps the model decide which other words matter most here.
A transformer is the architecture that puts this together.
Take:
The child sat by the bank and watched the ducks.
The embedding for bank includes several possible meanings: money, account, river, edge, building. Attention helps the model give more weight to sat, ducks and watched, making the riverbank meaning more likely. The transformer repeats this process through many layers, refining the model’s representation of the text each time.
‘Till next week.
Mr A 🦾
Help a colleague save time by sharing this newsletter; distributing these ideas helps a friend get home on time and keeps our energy focused on what matters most: great teaching.
Safety & Privacy Notice
The tools and workflows mentioned are intended for professional productivity and educational enhancement. Users must ensure that any AI implementation remains compliant with their local data protection regulations and institutional safeguarding policies.
Data Privacy: Do not enter personally identifiable information (PII), sensitive student records, or confidential institutional data into public AI models.
Verification Required: AI-generated content can be inaccurate, biased, or out of date. Always maintain a "human-in-the-loop" approach by reviewing and fact-checking all outputs before use.
Professional Judgement: These suggestions do not substitute for formal legal, clinical, or safeguarding advice. Final responsibility for accuracy and appropriateness remains with the professional user.
