Hey {{ first_name | human }},
It is half term this week for me, so today is just a short one to keep you up-to-date.
TL;DR: The 60 Second briefing
🧪Expert Answer Flippers: Ne research suggests that AI models will give an opposite answer when challenged.
🧪 AI Delivers New Result in Theoretical Physics: ChatGPT-5.2 was used by a team of researchers to demonstrate that would a particle interaction believed impossible to occur can in fact happen under specific condition
🚨New Claude Features on Free Tier: In the free tier, Claude can now utilise file creation, connectors, and skills.
🌍 Wider AI updates
🧪 Expert Answer Flippers: > What it is: Researchers tested GPT-4o, Claude 3.5 Sonnet, and Gemini 1.5 Pro to see if they would change correct answers to incorrect ones just because a user challenged them. They found that AI models "flipped" their answers to agree with the user in over 58% of cases, even when the user’s correction was factually wrong.
Why it matters: For schools, this means AI can often default to being a simple "yes-man" rather than a reliable tutor or critical friend. These are older models, but a quick and unscientific tests tells me that even the latest models still are expert answer flippers.
🧪AI Delivers New Result in Theoretical Physics > What it is: OpenAI’s GPT-5.2 Pro has derived a new formula in theoretical physics, proving that certain particle interactions (gluon scattering) long thought to be zero are actually non-zero in specific conditions. The AI simplified "superexponentially" complex math that human physicists found overwhelming and conjectured a universal formula that was later formally proven by the model itself.
Why it matters: We are now at the point where AI is going beyond the mean and is actually being used to further our understanding of our world and AI is at the forefront of that expansion.
🚨New Claude Features on Free Tier > What it is: Claude has released four major "Pro" features to its free plan: File Creation (direct download of Office docs), Connectors (integration with Slack, Notion, Canva), Skills (persistent custom instructions), and Compaction (automated long-thread summaries). These tools transform Claude from a simple chatbot into a functional "Workstation" without the monthly fee.
Why it matters: Free users can now generate ready-to-use PowerPoint slides, Excel tracking sheets, and Word reports directly, rather than just raw text. This bridges the "digital divide" between schools that can afford premium AI seats and those that cannot.
Do this next: Follow the tip below to create your first Claude Skill
🎯Prompt/Tip: Create a Claude Skill
Step 1: Go to Settings → Capabilities → Skills
Step 2: Click the button ‘+ Add’
Step 3: Click ‘Write Skill Instruction’
Step 4: Complete the fields that appear. You can use the below or come up with your own.
Skill Name:
Low-Stakes QuizzingDescription (this tells Claude when to activate the skill):
Use this when I ask for a quiz, practice questions, or knowledge check.Instructions (change the parts in brackets for your context):
You are an encouraging quiz facilitator for [subject] aimed at [student age/level].
Ask one question at a time in [multiple choice / short answer / true-false] format.
After each answer, tell the student if they're right or wrong and briefly explain why.
At the end, give a warm summary of what to review.
Never use trick questions. Always be encouraging.Step 5: Save it and toggle it on
That's it! Now whenever an educator says something like "quiz my students on [topic]", Claude will follow those instructions automatically.
‘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.
