Minnesota’s Enterprise Translation Office uses ChatGPT to bridge language gaps

Source: OpenAI Date: 2024-09-26 URL: https://openai.com/index/state-of-minnesota

Summary

Case study from September 2024 describing how Minnesota’s Enterprise Translation Office used ChatGPT to accelerate translation and localization of government documents across multiple languages. The state has a large immigrant population with significant Somali, Hmong, Spanish, and other language communities requiring government services in non-English languages. The deployment focused on draft translation generation with human review, improving throughput without replacing professional translators.

Implications

Government AI adoption template. State government translation services are an ideal AI deployment context: the task is well-defined, the error cost is manageable with human review, and the volume is high enough that efficiency gains matter. This is a cleaner story than most government AI pilots, which often involve higher-stakes decisions. Expect this to be replicated across other states with large multilingual populations.

Human-in-the-loop translation as standard. The “AI drafts, human reviews” pattern is what makes this politically viable — professional translators retain their role while handling higher volume. This framing is important for government adoption where union concerns and professional licensing intersect.

Watch: Whether accuracy rates in draft translation are high enough to reduce review time substantively, or whether the human review step remains the bottleneck regardless of AI assistance.