Communication across languages has always been a cornerstone of international cooperation. But translation’s not what it used to be. AI is revolutionizing cross-linguistic dialogue, with LLM-powered technology becoming a given in professional settings. The influx of AI into language tasks, mixed with the culturally nuanced nature of language surrounding geopolitical concepts, makes understanding the limits of LLMs relevant to institutions that rely on translation and host political dialogue. We need language operations policies that reflect today’s reality, where humans and technology both engage in meaning-making but have differing relationships with liability. Responsible AI-augmented translation would be human-centred, accounting for culturally specific mediation between linguistic groups.
Anna Legault provides actionable recommendations for procedural transparency in AI-supported language operations, human-centred augmented translation frameworks, and defining use-cases for automated translation based on importance.
