What the evidence says about adaptive learning, gamification, vocabulary apps and generative AI. This is the design brief for RS Insight, the RS Learning System, Coursebook-to-Lesson and the homeschool platform.
Where the sources disagree
- Gamification evidence: the British Council Indonesia evaluation reports teachers adopting gamification (self-report). Kerr and Kohn argue points and leaderboards give short-lived extrinsic motivation. Neither measured learning outcomes.
- AI and learning: Deng et al. (2024) found performance gains; Kerr says performance ≠ learning, and Stadler et al. (2024) found reduced mental effort and depth.
Design principles
- Solutionism - define the problem before the technology
- Adaptive learning - hype vs evidence
- Gamification as pointsification
- Vocabulary apps - reward design outruns learning design
AI
- Does ChatGPT enhance learning - performance is not learning
- Critical AI literacy is hard to teach
- AI guidelines for teachers - six principles
Motivation evidence
- Rewards undermine interest in the rewarded activity
- Against competitiveness and rankings
- Spaced retrieval is the best-supported technique
More notes (by source)
Adaptive Learning in ELT (Philip Kerr)
- AI for Good - read vendor-sponsored AI claims critically
- Edtech winners and losers - the product matters
- Educational neuroscience mostly confirms psychology - beware seductive details
- Metaphors of AI - tool, cultural technology, stochastic parrot
- Word-card apps are a first stage - add generation, elaboration and hints
After Octopus (Vedrana Vojkovic Estatiev) - Design peer review so it produces useful feedback
- Written correction is a workload and uptake problem
Blog de Cristina (Cristina Cabal) - Personality adjectives - false friends, word cloud, bingo, tendency language
- Simultaneous-response team games for grammar review
Book - The Anxious Generation (Jonathan Haidt) - Phone-based childhood - what it means for teen classrooms
ELT-CATION (Svetlana Kandybovich) - Lesson planning with AI - you outline, the model drafts options
- Turn writing feedback into a learning activity
Fluency First ELT (Anderson & McCutcheon) - Online task bank for teenagers
Hana Tichá (How I see it now) - AI in the classroom - adopt slowly and redesign assignments
My ELT Rambles (Joanna Malefaki) - Moving a course online - audit platform and materials first
Nik Peachey (Peachey Publications) - AI as an interactive guide - empathy mapping one question at a time
- Human teachers vs AI teachers - care, belief and the digital divide
- Staged prompt templates for AI-assisted course design
- Telegram as a low-bandwidth blended learning platform
From my books (highlights)
Key terms (DELTA M1)
adaptive learning · computer-adaptive testing (IRT) · gamification / pointsification · extrinsic vs intrinsic motivation · novelty effect · solutionism · dataveillance · AI literacy · desirable difficulties · performance vs learning
Open questions
- Design rule set (synthesis): reward delayed retrieval and production, not logins; separate adaptive testing claims from adaptive learning claims; attempt first, AI second; publish data use.