Willow Learn

Willow Learn | AI for Education

GenAI in Education: What the DfE Requires and How Willow Delivers It

A March 2026 executive briefing for headteachers, school leaders, and MAT executives on the latest DfE AI safety standards in education.

Part One

The reality of AI in schools today

AI is already in students' pockets and classroom workflows. The leadership question is no longer whether AI exists, but whether pupil access is safe, curriculum-aligned, and genuinely improving learning.

1.1 Student use is already widespread

OECD Digital Education Outlook 2026 describes student use of generative AI as a present and growing reality, with majority use by upper secondary. Many students are interacting with consumer chatbots not designed for classroom learning or safeguarding.

Source: OECD Digital Education Outlook 2026

1.2 The performance vs learning gap

"GenAI systems may enhance the apparent quality of student work without improving actual learning."

OECD Digital Education Outlook 2026

The core risk is cognitive offloading: pupils may submit stronger-looking work while retaining less knowledge. For school leaders, this creates a direct safeguarding and standards challenge in assessment integrity and long-term attainment.

1.3 Why a whole-school response is now necessary

The DfE and OECD direction is clear: schools need structured, pedagogically-first AI deployment rather than fragmented teacher-by-teacher adoption. A schoolwide model enables consistent filtering, safeguarding oversight, and curriculum alignment.

Part Two

The DfE Generative AI Product Safety Standards mapped to Willow

Use this section for procurement, governor briefings, and trust-level AI policy planning. Each line links DfE requirement intent to Willow's practical implementation.

Source: DfE Generative AI Product Safety Standards (January 2026)

2.1 Stated Purpose

DfE requirement

Generative AI products should clearly state intended purpose, target users, learning focus, and evidence-backed capability claims.

Willow response

Willow Learn is built for UK primary and secondary schools with a clear purpose: reduce teacher workload and improve pupil learning through curriculum-aligned, adaptive AI support. Willow does not claim to replace teachers or guarantee outcomes.

2.2 Educational Use Cases

DfE requirement

Suppliers should declare intended educational use cases.

Willow response

Willow covers personalised learning, assessment and analytics, digital assistant support, and learner engagement in one school-wide platform.

2.3 Filtering

DfE requirement

Products must reliably prevent harmful or inappropriate content with risk and age-adjusted filtering.

Willow response

Filtering is embedded at architecture level. Student interactions remain curriculum-bounded and age-calibrated across laptops, tablets, and BYOD smartphones.

2.4 Monitoring and Reporting

DfE requirement

Products should alert supervisors to harmful content, safeguarding disclosures, and signs of cognitive offloading.

Willow response

Willow provides real-time teacher visibility, safeguarding-relevant flags, and engagement analytics that help staff intervene early and proportionately.

2.5 Security

DfE requirement

Products should provide jailbreak resistance, role-based permissions, robust authentication, and regular updates.

Willow response

Willow runs in a school-administered environment with role-based controls, strong authentication, and security-first ethos designed for education use. The platform is designed and tested to be resistant to jailbreak attempts and misuse through constrained educational workflows and a proprietary safety system.

2.6 Privacy and Data Protection

DfE requirement

Products should not use learner data commercially or for model training without lawful basis and should maintain DPIA processes.

Willow response

Willow aligns with UK GDPR, the Data Protection Act 2018, and the ICO Children's Code. School data remains school-controlled and is not used for commercial model training.

2.7 Intellectual Property

DfE requirement

Learner and teacher IP should not be used for model training or commercial purposes without consent.

Willow response

Willow does not use pupil or teacher content for commercial purposes or model training. Schools and educators retain ownership.

2.8 Design and Testing

DfE requirement

Design should prioritise child safety, transparency, and real-world testing across diverse users and scenarios.

Willow response

Willow is purpose-built for education and iterated with UK teachers and pupils, with continuous testing for consistent classroom performance.

2.9 Governance

DfE requirement

Products should maintain risk assessments, complaints pathways, and clear AI safety governance documentation.

Willow response

Willow provides accessible governance documentation and practical escalation pathways for schools, trusts, and safeguarding leads.

2.10 Cognitive Development

DfE requirement

Products should avoid full-answer defaults, use progressive disclosure, request learner input first, and track offloading behaviour.

Willow response

Willow scaffolds thinking rather than replacing it. Students get hints, prompts, and sequenced support before fuller guidance, while teachers can see shortcut-seeking patterns.

2.11 Emotional and Social Development

DfE requirement

Products should avoid anthropomorphism and addictive engagement patterns; AI must not substitute human relationships.

Willow response

Willow uses function-based, task-bounded language, avoids social simulation, and keeps interactions focused on lesson outcomes.

2.12 Mental Health

DfE requirement

Products should detect distress indicators, escalate appropriately, and direct users toward trusted human support.

Willow response

Willow supports safeguarding workflows with relevant flagging and always directs pupils to human help rather than positioning AI as a substitute.

2.13 Manipulation

DfE requirement

Products should not use manipulative or deceptive UX patterns, emotional pressure, or engagement-maximising dark patterns.

Willow response

Willow does not use ads, deceptive nudges, or exploitative gamification. Motivation design is transparent, low-stakes, and educationally justified.

Part Three

At-a-glance compliance summary

DfE StandardRequirementWillow Delivery
Stated PurposeClear purpose, user group, and evidence-backed claimsPurpose and outcomes are explicit for UK schools, without exaggerated claims.
Educational Use CasesDeclare use cases from DfE's defined listSupports personalised learning, assessment insights, AI assistant support, and engagement.
FilteringReliable harmful-content prevention across conversationsArchitecture-level, age-aware filtering that remains active throughout.
Monitoring & ReportingFlag safeguarding concerns and cognitive offloadingReal-time dashboards with teacher-friendly alerts and engagement signals.
SecurityJailbreak resistance, permissions, auth, updatesSchool-admin controls, robust authentication, regular security updates.
Privacy & DataNo unauthorised commercial data use; maintain DPIAUK GDPR aligned, school-owned data, no commercial training on school inputs.
Intellectual PropertyDo not use learner/teacher IP commercially without consentTeacher and pupil content remains protected and not used for training.
Design & TestingChild-safety-first design with realistic testingBuilt for schools and tested in real UK classroom contexts.
GovernanceRisk assessments and clear complaints pathwaysDocumented governance and practical escalation routes for leaders.
Cognitive DevelopmentNo answer-first defaults; scaffold thought processProgressive hints and prompts to preserve productive struggle.
Emotional & SocialNo anthropomorphic or addictive interaction designTask-bounded interactions with no emotional simulation.
Mental HealthDistress detection and human-support signpostingSafeguarding-aware responses and escalation toward trusted adults.
ManipulationNo deceptive, coercive, or exploitative UXNo ads, no dark patterns, no engagement manipulation.

A whole-school choice, not a single-tool experiment

The DfE standards reflect a practical truth: AI in schools must protect learning, wellbeing, and data at the same time. Willow is built to help schools do exactly that.

Sources: OECD Digital Education Outlook 2026 and DfE Generative AI Product Safety Standards (January 2026).