Education · 17 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 and Standards

DfE Standards Shift Responsibility to Vendors

New guidance on Generative AI Product Safety Standards marks a significant shift, moving the burden of safety from schools to tech providers. The standards require filtering to be embedded within products rather than bolted on, ensuring systems understand pedagogical context rather than just blocking keywords. This comes as Google's Gemini 3 infrastructure begins entering the edtech ecosystem.

Why it matters: Schools can now demand "safety by design" in procurement rather than relying solely on local network filtering to catch AI outputs.

Read the source on LinkedIn →

41% of Teachers Report No Agreed AI Approach

New research from Computing At School (CAS) reveals a concerning policy gap, with 41% of teachers stating their school lacks an agreed approach to AI and a further 17% unsure if one exists. The data suggests this is causing uncertainty rather than resistance, with many staff using tools in "stealth mode" to avoid potential professional risk.

Why it matters: Silence from leadership is being interpreted as a lack of permission, preventing the sharing of best practice and workload-reducing strategies.

Read the source on LinkedIn →

Data and Safeguarding

Potential Data Breach Alert for ChatGPT Edu

Researchers have identified a potential vulnerability in ChatGPT Edu implementations that may affect student data privacy. While specific details are being compiled, early warnings suggest institutions using this specific tier need to review their data isolation configurations immediately. This serves as a stark reminder that even "enterprise" agreements require rigorous penetration testing.

Why it matters: Data controllers in MATs should pause any new rollouts of this specific tool until the vulnerability is patched or clarified.

Read the source on LinkedIn →

New Safeguarding Guide for Safer Internet Day

To coincide with Safer Internet Day, NetSupport has released an updated safeguarding guide specifically addressing the 2026 digital landscape. The resource covers the roles of DSLs, governors and teachers in managing risks associated with deepfakes and AI-generated content. It includes specific checklists for auditing school compliance against the latest statutory guidance.

Why it matters: This provides a ready-made framework for your spring term safeguarding audit regarding AI risks.

Read the source on LinkedIn →

Strategy and Market

The "AI Bubble" and School Procurement

Financial analysts at the FT are warning of a market correction in the AI sector, describing it as "eating itself" due to high costs and low conversion to paid subscriptions. For education, this poses a sustainability risk for schools signing long-term contracts with venture-backed startups that may not survive the correction.

Why it matters: Due diligence must now include financial health checks of suppliers to ensure tools introduced today will still exist next academic year.

Read the source on LinkedIn →

Criticism Mounts Over AI Tutoring for Disadvantaged Pupils

As the government's £23m AI pilot rollout continues, experts are raising concerns about relying on automated tutoring for disadvantaged students. Critics argue that AI cannot replicate the emotional regulation, motivation and "world knowledge" provided by human teachers, potentially widening the attainment gap if used as a replacement rather than a supplement.

Why it matters: Schools participating in the pilot should ensure AI tutors are used to support, not replace, human intervention for vulnerable learners.

Read the source on LinkedIn →

On your radar

Context-Aware FilteringThe new product safety standards require filters to understand pedagogical intent. Update your IT policy to differentiate between a student searching for "weapons" for a history project versus a safety threat.
The "Shadow Policy" RiskWith 17% of staff unsure if an AI policy exists, simply having a document on the server is insufficient. Re-launch your acceptable use policy in the next staff briefing to ensure clarity and reduce "stealth" usage.
Protecting "Hard Tasks"Identify specific assignments that must remain analogue to build resilience. Tasks like handwritten synthesis of multiple sources or real-time face-to-face debate should be explicitly ringfenced from AI assistance in your curriculum planning.
Vendor Financial HealthGiven the predicted "tech wreck" in the AI sector, add a clause to your risk register regarding software continuity. Avoid locking into multi-year contracts with early-stage startups without a data exit plan.
Cognitive SovereigntyFollowing Sweden's move to reintroduce paper-based learning, review your balance of digital vs. physical tasks. Ensure students are developing the "cognitive seeds" of judgment and logic before they are given tools that outsource those processes.
NewerAI in Education, 24 February 2026OlderAI in Education, 10 February 2026