Hey {{ first_name | Colleague }},
This week’s briefing moves from "permission" to compliance, as UK regulators draw firm lines around safety and assessment standards
TL;DR: The 60 Second briefing
🚨UK AI Safety Updates: The DfE has published an update on its safety guidance around the use of AI in schools.
🚨Humans still key in marking: The chair of the Office for Qualifications and Examination Regulation does not think AI is ready to mark high-stake assessments
🧪Anthropic Index Update: Claude speeds up more complex tasks than the routine ones. Educational use is on the rise.
📚 AI+education news
🚨 Safety Guidance Update 2026 > What it is: The new DfE standards mandate that educational AI must use "progressive disclosure" (hints instead of answers) and implement "hard limits" on usage to prevent cognitive offloading. Products must also detect "learner distress" and relationship formation with the AI, flagging these directly to the school’s Designated Safeguarding Lead (DSL).
Why this matters: This provides a statutory "Procurement Shield". You can now demand that any AI vendor proves compliance with these standards. In terms of learning, by design, the student needs to do the cognitive heavy lifting rather than offloading it to the machine.
Do this next: Ask your current AI providers for their compliance statement against the Jan 2026 "Generative AI: Product Safety Standards".
🚨Humans still key in high-stakes marking > What it is: Ofqual chief Sir Ian Bauckham clarified that AI is not yet ready to replace human markers for national qualifications. He cited significant "difficulties" AI must overcome, including "anxiety" over its use in A-level English and History coursework.
Why it matters for schools: This is a firm regulatory boundary. While AI can assist with feedback, schools cannot use it as the final arbiter for "high-stakes" grades without risking compliance.
Do this next: If your department/school uses AI for "mock" marking, ensure you have a "Human Sign-off" step. Every AI-generated grade must be verified by a teacher to stay aligned with Ofqual’s stance.
🌍 Wider AI updates
🧪 Anthropic Index - AI for "Complex Work" > What this is: Anthropic's new "Economic Index" (Jan 2026) reveals that AI is boosting complex work faster than routine tasks. Tasks requiring a university-level education see a 12x speedup, compared to 9x for high-school-level tasks. Educational activities (including tutoring and content generation) accounted for 15% of all interactions.
Why it matters for schools: This debunks the idea that AI is only for "simple" admin. It shows that the more knowledge-intensive a task (like curriculum design or data analysis), the higher the potential time-save.
Do this next: Don't just use AI to reply to emails. Find a high-value task that you could invite AI to the table to and see how it can be used to help synthesise that information.
🚨 The "Ad-Supported" Classroom?> What this is: OpenAI has announced it will start testing advertisements in the free and "Go" tiers of ChatGPT in the US. They have promised not to bring these in for education accounts.
Why it matters for schools: While OpenAI pledges ads won't influence answers, they could introduce "distractive load" for students if you rely on the free version.
Do this next: Those outside the US just need to monitor the ad role out.hose in the US will want to consider if th If ads become the default on the free tier, then you may need to discuss upgrading to the educational accounts or move to a different LLM.
🎯Prompt: The ‘Expert-in-the-Loop’ Marking Auditor
I am a [SUBJECT] teacher conducting a formal internal assessment. Ofqual has mandated that while technology can assist in assessment, a "human-in-the-loop" is required to ensure marking remains accurate, fair, and free from "black box" errors for high-stakes exams.
Act as a Senior Examiner and Professional Auditor. Your objective is to provide the initial professional feedback and marking suggestions that a human teacher will then validate or correct based on official exam board criteria.
Compare the provided [STUDENT WORK] against the [OFFICIAL MARK SCHEME].
ANALYSE the student's work for specific evidence that meets the mark scheme descriptors.
GENERATE a suggested mark and a "Professional Justification" note that explicitly links the student's text to the marking criteria.
HIGHLIGHT any areas of the work that are borderline or require a specific human professional judgment call.
Specifically, look for [e.g., specific skills like "analysis vs description" or "AO1 vs AO2" weighting]. Ensure you prioritise accuracy over fluency. [PASTE STUDENT WORK HERE].
Format the output as a table with three columns: 1) Marking Criteria, 2) Evidence found in the work, and 3) Suggested Mark/Feedback. Use a formal, objective tone. Do not include student names or PII.
Before providing the final output, ask me if there are any specific examiner "booster" documents or additional mark scheme clarifications I want you to consider.Before you or your staff upload a single essay to an AI tool, ensure you are compliant with the regulations around uploading student data in your area.
‘Till next week.
Mr A 🦾
Help a colleague save time & keep up-to-date 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
