GDST AI Adoption — "Learn Fast, Act Slowly"
Research Date: 2 March 2026 Relevance to GenEvolve: The most thoughtful AI adoption framework in UK education. Shelley specifically referenced Guildford High School's AI approach. Provides the blueprint for how GenEvolve should integrate AI — values-first, teacher-confident, student-critical.
Organisation Overview
| Field | Detail |
|---|---|
| Full Name | Girls' Day School Trust (GDST) |
| Type | UK's leading network of independent girls' schools |
| Schools | 23 schools + 2 academies |
| Students | ~20,000 |
| Founded | 1872 |
| Headquarters | London |
| Key Person | Rachel Evans — Director of Digital Transformation |
| AI Philosophy | "Learn Fast, Act Slowly" |
| Influenced by | Professor Rose Luckin (UCL Knowledge Lab) |
| Website | gdst.net |
Sources: GDST website, GDST SLT Conference 2025
Rachel Evans — Director of Digital Transformation
Rachel Evans is the architect of GDST's AI strategy and one of the most influential voices on AI in UK education:
| Detail | Information |
|---|---|
| Role | Director of Digital Transformation, GDST |
| TEDx Talk | October 2024 — feminist perspective on AI |
| SLT Conference | February 2025 — "Redefining Technology in Education" presentation |
| Philosophy | Values-led, ethics-first, feminist approach to AI |
| Key Quote | Describes the AI landscape as "intensely masculine and aggressive" |
| Mission | Equipping young women with confidence, creativity, and ethical foundations for AI |
Evans advocates for girls to be creators and shapers of AI, not just consumers — directly challenging the gender gap in technology.
Sources: Rachel Evans TEDx, GDST Executive Team
The "Learn Fast, Act Slowly" Framework
Origin
- Coined by Rachel Evans
- Inspired by Professor Rose Luckin (UCL Knowledge Lab) — world-leading AI-in-education researcher
- Core principle: develop deep understanding of AI's capabilities and limitations before embedding in curriculum
What It Means in Practice
| Phase | Approach |
|---|---|
| Learn Fast | Aggressively explore AI tools; run pilots; train teachers; experiment in controlled settings |
| Act Slowly | Don't deploy across the network until you understand implications for: learning outcomes, academic integrity, data protection, equity, wellbeing |
Critical Distinction
This is NOT a "wait and see" approach. It's "understand deeply, then deploy deliberately." The distinction matters enormously — GDST is not behind on AI; it's ahead on responsible AI.
AI Tools in Use (2025-2026)
For Teachers
| Tool | Context | Data Protection |
|---|---|---|
| Microsoft Copilot | Productivity, lesson planning, admin tasks | Enterprise accounts (data protected) |
| Google Gemini | Research assistance, content generation | Enterprise accounts (data protected) |
Key design choice: Teachers access AI through enterprise systems with data protection guarantees — not personal accounts. This is a non-negotiable for GenEvolve.
For Students
| Tool | Context | Safeguards |
|---|---|---|
| ChatGPT | Critical analysis exercises, exposure | Controlled educational context only |
| Figma | Design projects with AI features | Student use with safeguards |
| Canva | Creative projects with AI features | Student use with safeguards |
Students interact with AI in controlled, educational contexts — never unsupervised, never with personal accounts.
Why Girls Are Skeptical of AI — Critical Research
The Trust Problem
GDST has observed that their students are more skeptical of AI than many adults:
| Pattern | Detail |
|---|---|
| Accuracy concerns | Students encounter AI hallucinations — plausible but incorrect answers |
| Cheating anxiety | Many girls worry using AI constitutes "cheating" — want work to reflect their own thinking |
| Training barrier | Some feel they need extensive training before engaging with AI |
| Critical consumption | GDST encourages this skepticism; teaches questioning over acceptance |
Specific AI Errors
Royal High School Bath documented that "closer inspection reveals the limitations of the program and its makers have admitted that it 'sometimes writes plausible-sounding but incorrect or nonsensical answers.'"
GenEvolve Insight
This skepticism is a strength, not a weakness. GenEvolve's AI features should: - Surface uncertainty explicitly ("I'm not confident about this answer") - Include "verify this" prompts - Reward questioning and critical thinking - Never position AI as infallible
"Responsible AI Use" Policy Framework
Five Pillars
| Pillar | Description |
|---|---|
| Ethics first | AI decisions filtered through Trust values, not efficiency gains |
| Sustainability | Technology choices align with environmental/social sustainability |
| Human-centred | AI amplifies human capabilities; doesn't replace creativity or emotional intelligence |
| Critical consumers | Staff and students question technology's implications |
| "Whether" before "how" | The question isn't "how to use AI" but "should we use AI for this specific purpose" |
The "Whether" Question — Strategic Gold
For each AI feature proposed, GDST asks: Should we use AI for this? Not just "how." This forces every AI feature to justify its existence through values.
GenEvolve version: For every AI feature (tutoring, assessment, wellbeing monitoring, content generation), run a values-based assessment: - Does this serve the student's growth? - Does it respect their autonomy? - Does it protect their data? - Could it cause harm? - Does it pass the "whether" test?
Guildford High School — Shelley's Specific Reference
Shelley Crowther specifically referenced Guildford High School's AI experience:
What They've Done
- AI Literacy sessions — dedicated programme for Year 10 and Lower Sixth (ages 14-17)
- Part of academic enrichment programme
- iPads and Apple Pencils throughout the school
- AI woven into creative, collaborative, and analytical work
- Focus: preparing girls to "lead in an AI-powered future"
Shelley's Key Takeaways
- Girls need to see AI as a tool they control, not a crutch they depend on
- Trust must be earned through transparency about limitations
- AI literacy is as important as digital literacy
- The wellbeing dimension (anxiety, comparison, dependency) must be addressed proactively
Curriculum Integration Across GDST
AI is not a standalone subject but woven across disciplines:
| Subject | AI Integration |
|---|---|
| PSHE | Online safety, AI and digital citizenship |
| English | AI's societal impact, analysis of AI-generated text |
| Civil Discourse | Ethical debates about AI |
| Computer Science | Technical understanding of how AI works |
| Design | AI features in Figma (with safeguards) |
| Creative Arts | AI features in Canva (with safeguards) |
| Critical Thinking | Evaluating AI outputs, source verification |
The Gender Dimension
Rachel Evans explicitly addresses the gender gap in AI:
| Challenge | GDST Response |
|---|---|
| AI field is "intensely masculine and aggressive" | Prepare girls with confidence, creativity, ethical foundations |
| Girls may approach AI more cautiously | Lean into this as a strength — critical thinking over blind adoption |
| Gender gap in technology careers | Position girls as future AI leaders, not just users |
| Authenticity concerns | Validate desire for work to reflect own thinking |
GenEvolve application: If serving any demographic that approaches AI with caution (girls, neurodivergent students, anxious learners), design AI interactions that reward curiosity and critical thinking, not just efficiency.
Strengths
- Most thoughtful AI framework in UK education
- Values-led — ethics and sustainability embedded, not bolted on
- Enterprise data protection — AI via secure enterprise accounts only
- Teacher-first — teachers confident before students engage
- "Whether" before "how" — strategic gold for responsible AI adoption
- Rose Luckin influence — world-leading AI-in-education research informing practice
- Rachel Evans leadership — articulate, visible, credible advocate
- Curriculum-wide integration — AI across all subjects, not just Computer Science
- Gender-aware — explicitly addressing how girls interact with AI
- Guildford High School — specific case study Shelley references
Weaknesses
- Girls' schools only — framework untested in co-ed or boys' contexts
- Traditional schools — still operate within exam-based system
- Independent sector — £15,000+/year fees; not accessible to all
- Conservative pace — "act slowly" may mean slower adoption than competitors
- No alternative assessment — AI supports exam preparation, not portfolio assessment
- No SEND focus — not specifically designed for neurodivergent learners
- No technology platform — GDST uses third-party tools, doesn't build its own
GenEvolve AI Integration Checklist (Based on GDST Lessons)
| Step | Action |
|---|---|
| 1 | Teacher AI literacy first — every educator trained before student-facing AI launches |
| 2 | Enterprise-grade data protection — AI through GenEvolve's infrastructure, not third-party personal accounts |
| 3 | "Whether" test — every AI feature justified through values assessment |
| 4 | Build for healthy skepticism — surface uncertainty, include "verify this" prompts |
| 5 | Wellbeing dimension — monitor for AI dependency, frustration, disengagement |
| 6 | SEND-specific AI design — AI adapted for neurodivergent interaction patterns |
| 7 | Gradual rollout — pilot with small cohort, iterate on real data, then expand |
| 8 | Cross-curriculum integration — AI literacy in every subject, not siloed |
| 9 | Parent transparency — show parents exactly when/how AI is used, what data it accesses |
| 10 | "Learn fast, act slowly" — adopt as GenEvolve's official AI integration philosophy |
Strategic Assessment
The GDST's AI approach is the gold standard for responsible AI in education. For GenEvolve — building an AI-native platform — the lessons are non-negotiable:
- AI literacy is curriculum, not technology — it belongs in every subject
- Trust is earned through transparency — show limitations, not just capabilities
- Enterprise security is non-negotiable — especially for SEND students
- The "whether" question prevents feature bloat — not every AI feature is worth building
Recommendation: Adopt "learn fast, act slowly" as GenEvolve's official AI philosophy. Reference GDST in conversations with parents and investors — it signals thoughtfulness and alignment with the UK's most respected educational networks.
Threat Level: NONE (not a competitor; the most important strategic reference for GenEvolve's AI approach)
Sources: GDST website, Rachel Evans TEDx, Wimbledon High School blog, Royal High School Bath, Guildford High School, Shelley Crowther transcript (27 Feb 2026). Cost: Gemini grounding (free tier).