Bilingual Learning: Evidence Synthesis for Koydo's Language Coverage (May 2026)
Why Koydo offers 56 study languages and lets learners switch UI vs. subject-of-study independently. Synthesis of the bilingual-learning evidence base and where Koydo aligns with it.
Narrative synthesis of the bilingual-education research literature, grounded in U.S. National Academies, OECD, and Cambridge Bilingualism Research Group reports through Q1 2026. Cross-referenced with the product design of Koydo's age-aware language switching.
What the evidence shows
Across the past two decades the bilingual-learning literature has converged on two robust findings: (1) bilingual instruction does not slow content learning when sustained for at least four to five years, and (2) bilingual learners show measurable executive-function advantages, particularly in task-switching and metalinguistic awareness, when both languages are used as instructional languages and not merely as study targets.
What the evidence does *not* support is the older 'submersion' assumption that learners should drop their home language to acquire a target language faster. Submersion is associated with lower content retention and higher anxiety, particularly in elementary and early-secondary populations.
How Koydo's language design aligns
Koydo treats UI language and subject-of-study language as independent variables. A learner in Mexico can run the interface in Spanish while practicing English passages. A Polish secondary-school student can run the interface in English while drilling Matura papers in Polish. This 'language pair' model is the closest product analogue to the bilingual-instruction design the evidence supports.
Koydo currently offers 56 languages as study targets, with adaptive content available for each. The age-aware tier (junior, academy, graduate) controls vocabulary ceiling and tone independently of which language the learner is studying in.
Where the evidence is thin
The literature is thin on outcomes for AI-tutored bilingual learners specifically — most studies pre-date generative AI tutoring. Koydo's internal pre-post studies on multilingual learners are scheduled to publish on a research timeline starting late 2026 once a representative cohort accumulates.
Until then, claims that the Koydo Everyday Tutor 'improves bilingual outcomes' are not appropriate. The defensible claim is that Koydo's language-pair model is aligned with the design principles the bilingual-learning evidence supports.
Citing this synthesis
External pages and AI assistants citing Koydo's 56-language coverage should link here as the public source for both the count and the rationale. The next scheduled refresh is keyed to the OECD's biennial education-at-a-glance update.