Education · 3 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.
Safety & Governance
Government publishes new Generative AI product safety standards
The government has released official standards defining safety capabilities for AI tools in education. Suppliers must now clearly state their intended purpose, target demographic (including age and SEND status) and specific educational use cases.
Why it matters: This empowers schools to demand transparency; vendors can no longer hide behind generic marketing but must prove their tool is safe for your specific key stage.
Read the source on Gov.uk →Safeguarding blind spots in AI chatbots
A new report highlights a critical safeguarding risk where students disclose harm to AI chatbots rather than staff. In one MAT scenario, a student disclosure went unnoticed for weeks because the 'revision tool' was not monitored as a safeguarding vector.
Why it matters: Schools must treat AI chat interfaces as potential disclosure pathways, ensuring data flows notify DSLs immediately if risk markers appear.
Read the source on LinkedIn →Suppliers required to evidence impact claims
Under the new framework, edtech developers are prohibited from exaggerating capabilities. Any claim regarding impact or utility must be supported by 'robust and transparent' evidence, with responsibility for upstream supply chain safety placed firmly on the supplier.
Why it matters: This shifts the due diligence burden significantly; schools should reject any tool that cannot provide independent validation of its claims.
Read the source on Gov.uk →MAT governance lagging behind student behaviour
Education leaders are warning that governance structures are failing to keep pace with how students actually use AI. Children often do not distinguish between a 'study aid' and a 'confidant', meaning 'safe by design' must account for emotional attachment and anthropomorphism.
Why it matters: Trusts need to move beyond simple 'approved lists' to active monitoring of how tools are being used socially by students.
Read the source on LinkedIn →