Places in Language Departments in Korean Higher Education Before and After ChatGPT, 2016-2026: A Descriptive Note

A common expectation is that capable machine translation and large language models reduce student demand for foreign-language study. This note describes places offered and applications received by language departments in the full department-level admissions file published by the Korean Ministry of Education for every higher-education institution, 2011 to 2026 (317,154 department-year rows), with the 2016-2022 intakes before ChatGPT and the 2024-2026 intakes after it (the 2023 intake straddles the release). In-quota places in foreign-language departments fell from 28,324 in 2016 to 16,466 in 2026 (-41.9%): by 27.3% between 2016 and 2022 and by a further 20.1% between 2022 and 2026. Places in other humanities fell by 35.6% over 2016-2026 and places in all subfields outside the studied groups by 0.3%. Places in departments whose names contain interpreting or translation were stable to 2022 (474 in 2016, 487 in 2022) and then fell to 292 at the 2023 intake and 181 in 2026 (-62.8% from 2022). Applicants to foreign-language departments fell by 33.9% over 2016-2026, less than places. The analysis plan's primary outcome, applicants per place relative to other humanities, is not used for foreign languages, because a stop condition in the analysis plan (unpublished; its timing cannot be verified) fired: applicants on rows with zero places exceeded 5% of the group's applicants in one year (5.59% in 2015). The value the plan would have reported is listed in an appendix as exploratory and supports no conclusion. For AI departments no before-and-after comparison is possible: the AI engineering subfield exists only from 2022, and name-matched AI and data departments grew from 300 to 10,570 places, 65.4% of that increase by the 2022 intake. The file shows when places changed; it does not show why.