Links to claims

Supports (1)

Claim	Basis
K16 (candidate)	Note carries the bank kill sentence behind K16 (group K16); no change in review

Claim

Basis

Note carries the bank kill sentence behind K16 (group K16); no change in review

Challenges (1)

Claim	Basis
C2	Positive pooled effect of ChatGPT; the bank notes that without closed-book outcomes this is performance, not learning. Open test for C2.

Claim

Basis

Positive pooled effect of ChatGPT; the bank notes that without closed-book outcomes this is performance, not learning. Open test for C2.

Review state: first-pass draft, pending human review.

Source note (Swedish, verbatim from the Kunskapsbank)

Syntes: Deng m.fl. (2025) — Does ChatGPT enhance student learning?

ID: AI-005
Tier: meta
Closed-book follow-up: nej / blandat — författarna flaggar själva mätbias
HE vs K–12: mest HE; ~14 % K–12
Novelty decay: ja — tidiga studier riskerar novelty inflation
Utfallslabel: prestation, inte lärande

Metadata

ID: AI-005
Tier: meta
N: 69 artiklar; 62 i meta
Population: övergripande (HE-tung)
Country: blandad
Subject: ChatGPT student learning
Duration: varierar

Finding

SR+meta positiv snitteffekt; Deng et al. varnar för measurement bias. Utan closed-book: prestation, inte lärande.

Mechanism

Ospecificerad pool — inkluderar olikartade pedagogiska roller.

Limits

K–12-tunn; closed-book sällan; novelty.

This claim dies if …

Bias-korrigerad analys med endast delayed closed-book ger noll/negativ effekt.

Swedish implication

Measure med samma ribba som Bastani (closed-book); annars label prestation.

Does ChatGPT enhance student learning? A systematic review and meta-analysis of experimental studies