Do AI chatbots improve students learning outcomes? Evidence from a meta-analysis
Lower secondary (grades 7–9) Upper secondary Across ages Adults / lifelong learning AI and learning
Links to claims
Supports (1)
| Claim | Basis |
|---|---|
| K30 (candidate) | Note carries the bank kill sentence behind K30 (group K30); no change in review |
Challenges (1)
| Claim | Basis |
|---|---|
| C2 | Positive chatbot effects, stronger in higher education than K–12; closed-book rare. Open test for C2. |
Review state: first-pass draft, pending human review.
Source note (Swedish, verbatim from the Kunskapsbank)
Syntes: Wu & Yu (2024) — AI-chatbots meta (tutoring systems)
ID: AI-007
Tier: meta
Closed-book follow-up: nej / sällan rapporterat som delayed without-bot
HE vs K–12: HE > K–12
Novelty decay: ja — moderator: novelty decay över tid
Utfallslabel: prestation, inte lärande
Metadata
- ID:
AI-007 - Tier: meta
- N: 24 randomiserade studier
- Population: HE > K–12
- Country: blandad
- Subject: AI chatbots learning outcomes
- Duration: ofta kort
Finding
Meta positiva chatbot-effekter; starkare i HE än K–12; novelty moderator. Closed-book: nej/sällan → prestation, inte lärande om botten finns kvar vid test.
Mechanism
Chatbot som scaffold/feedback; risk att effekten är tillgång till svar snarare än encoding.
Limits
PAYWALL fulltext; HE-dominans; novelty.
This claim dies if …
Effekten försvinner i K–12 closed-book efter novelty-fönstret.
Swedish implication
Stop extrapolera HE-chatbot-g till svensk grundskola; measure novelty-vecka 1 vs delayed closed-book.
Links
- DOI: https://doi.org/10.1111/bjet.13334
- Relaterade: AI-002, AI-004, AI-005, AI-003
Metadata as recorded
- ID
- AI-007
- Tier
- meta
- Closed-book follow-up
- nej / sällan
- HE vs K–12
- HE > K–12
- Titel
- Do AI chatbots improve students learning outcomes? Evidence from a meta-analysis
- Författare
- Wu, R.; Yu, Z.
- År
- 2024 (online 2023)
- Källa/DOI
- https://doi.org/10.1111/bjet.13334
- URL
- https://bera-journals.onlinelibrary.wiley.com/doi/abs/10.1111/bjet.13334
- Typ
- meta-analys (24 randomiserade studier)
- Åldersgrupp
- övergripande (HE > primary/secondary)
- Teman
- ai-och-larande (tutoring systems / chatbots)
- Sparade fil
- ABSTRACT.md (full PDF paywall Wiley/BJET vid hämtning)
- Hämtad
- 2026-09-07
- Paywall
- JA — köp/institutionell åtkomst via Wiley; ERIC EJ1408598 har abstrakt
Original files held in the project (not published): ABSTRACT.md