Pedro Llanos-Paredes and Carl Benedikt Frey

Rapid advancements in artificial intelligence (AI) have sparked debate over its employment effects, yet evidence on AI’s labor market impacts remains scant. This study investigates the labor market effects of machine translation (MT) on a) employment and wages in the translation profession; and b) the demand for foreign language skills across occupations and industries. Taking advantage of the heterogeneity in the use of MT across 695 local labor markets in the United States, we analyze its effects post-2010, when Google Translate was released as an app. Doing so, we document a negative relationship between Google Translate adoption and translator employment, corroborated by an instrumental variable approach, and a host of placebo regressions. Similarly, improvements in MT reduced the demand for all foreign language skills investigated, including for Spanish, Chinese, Japanese, French, and German.