Education · 10 February 2026

AI in Education

For teachers, school leaders and MATs

At askKira, we provide GDPR-compliant, secure AI built specifically for UK schools. Developed with over 100 education experts and aligned with DfE, Ofsted and EEF guidance, askKira helps teachers, school leaders and MATs work smarter, reduce workload and make data-driven decisions with confidence.

This week's roundup brings you the latest AI developments that matter for UK education.

Policy & Funding

DfE expands EdTech Testbed with £23m injection

The Department for Education is actively recruiting schools for the largest evidence-based study into AI in the classroom. The programme has received a £23 million funding boost to determine what actually works in UK settings. The focus is shifting from automation for its own sake to individualised learning at scale, ensuring tools amplify rather than replace teacher judgement.

Why it matters: This signals a move from "wild west" adoption to rigorous evidence gathering, offering schools a chance to pilot tools with central support before committing budgets.

Read the source on LinkedIn →

Schools pause Gemini and Copilot over safety standards

Schools using Google and Microsoft ecosystems are facing compliance hurdles following new government safety standards. Reports indicate that standard versions of Gemini and Copilot do not currently meet specific DfE requirements, such as mandatory "graduated disclosure" (guiding students rather than giving answers) and the prohibition of anthropomorphic "I" statements.

Why it matters: IT leads must urgently review if their current enterprise licences cover these specific educational safeguards or if student access needs to be suspended pending updates.

Read the source on LinkedIn →

Research & Evidence

OECD report reveals the 'competency trap' of AI use

New data from the OECD suggests a complex relationship between AI use and academic performance. While students using GenAI for open-book tasks performed better, their performance in closed-book exams dropped by 17% compared to those who studied without AI assistance. The report highlights that efficiency in completing tasks does not equate to efficacy in retaining knowledge.

Why it matters: Schools need to audit their homework policies to ensure AI is used for "productive struggle" rather than bypassing the cognitive effort required for memory retention.

Read the source on LinkedIn →

Brookings Institution warns of screen time correlation

A new policy paper from the EDSAFE AI Alliance and Brookings Institution argues that current AI integration often mimics passive screen time rather than active learning. The research indicates that as digital consumption increases, executive function and critical thinking skills in younger students may decline if the technology replaces foundational human interaction.

Why it matters: Primary leaders should be wary of "AI tutors" that remove the social aspect of learning, prioritising hands-on problem solving before digital integration.

Read the source on LinkedIn →

Safeguarding

Snapchat 'My AI' and emotional dependency risks

As Safer Internet Day approaches, new surveys from Internet Matters reveal that over a third of children aged 9-17 feel their AI chatbot interactions resemble a friendship. Vulnerable students are increasingly disclosing feelings to bots like Snapchat's "My AI" rather than trusted adults. These systems cannot detect distress or escalate safeguarding concerns effectively.

Why it matters: DSLs must update safeguarding training to cover "emotional anthropomorphism" and ensure parents know how to disable AI features in social apps.

Read the source on LinkedIn →

On your radar

Safer Internet Day 2026With the theme "Smart tech, safe choices" occurring this week (10 Feb), use assembly time to specifically address the difference between an AI "assistant" and a "friend".
Oral Assessment AuditsFollowing the OECD findings, consider using AI to transcribe and analyse oral assessments rather than written homework. This allows teachers to spot reasoning gaps that students might hide behind AI-generated written text.
Metacognitive PlanningUpdate student prompts to focus on the "messy middle" of learning. Instruct students to use AI for sequencing and outlining a project (metacognition) but ban it for the final drafting phase to protect writing resilience.
Cognitive Protocols CheckIf your MAT uses Microsoft Copilot or Google Gemini, verify if your instance has "graduated disclosure" enabled. If the AI provides direct answers to academic questions without scaffolding, it may now be non-compliant with DfE safety standards.
Revision StrategyEncourage students to use tools like NotebookLM to organise their own "messy notes" into revision quizzes rather than generating content from scratch. This keeps the student's own work as the primary source material.
NewerAI in Education, 17 February 2026OlderAI in Education, 3 February 2026