English-delivered AI explanations impose a double cognitive burden: foreign-language content significantly reduced learning outcomes versus native-language instruction in controlled higher-education studies¶

Claim · All claims
Evidence · 1 study (2 entries) · review r2 · q3
Currency · newest study 2026 · checked against its sources 2026-10-18 (automated, 2 of 2 entries) · not reviewed by a person

Subclaims¶

q3 i? Exposure to academic content in a foreign language without explicit language support significantly reduced both content and language learning outcomes compared to native-language instruction (Roussel et al., 294 students). → Avijit Roy 2026 q2 i? Bilingual (English and Tamil) instruction outperformed English-only instruction for non-English-dominant students on linked-list concept learning in programming education. → Avijit Roy 2026 (2)

Evidence¶

Avijit Roy 2026¶

Avijit Roy, Proma Roy. (2026). Structural Silence: When AI Infrastructure Fails Speakers of Underrepresented Languages — A Case Study in Bengali, the Digital Divide, and the Hidden Cost of English-Centric Design. arXiv preprint. https://arxiv.org/abs/2608.12278 doi:10.48550/arxiv.2608.12278

q3 · i? · theoretical · r2

This article reports, citing Roussel et al. (2017), a controlled experiment with 294 higher education students in which foreign-language presentation significantly lowered content and language learning outcomes; no effect size is printed here.

"Roussel et al. (2017) demonstrated this experimentally with 294 higher education students, showing that content presented in a foreign language produced significantly lower outcomes across both content and language learning compared to native-language presentation, under conditions where all other instructional variables were held constant."

Avijit Roy 2026 (2)¶

Avijit Roy, Proma Roy. (2026). Structural Silence: When AI Infrastructure Fails Speakers of Underrepresented Languages — A Case Study in Bengali, the Digital Divide, and the Hidden Cost of English-Centric Design. arXiv preprint. https://arxiv.org/abs/2608.12278

q3 · i? · review · r2

This article reports, citing Soosai Raj et al. (2018), that bilingual English–Tamil instruction outperformed English-only instruction for non-English-dominant students; no effect size is printed here.

"Soosai Raj et al. (2018) found analogous patterns in programming education, with bilingual instruction (English and Tamil) outperforming English-only instruction for non-English-dominant students on linked-list concept learning."

Discussion¶

Related Claims¶

• In one participatory budgeting case, Latino immigrant participation lapsed under outreach with interpretation and was sustained only after the activity structure was redesigned — related
• English learners lag behind their peers in achievement and attainment; limited exposure to academic content is offered as a partial explanation, not tested — related