Principled Caution: Navigating AI in Asian Newsrooms
Independent newsrooms across eight Asian countries are adopting AI with "principled caution" due to risks of legal and political persecution from errors. A report by Asia Centre, DW Akademie, and the German Federal Ministry for Economic Cooperation and Development highlights this careful approach. The study, covering Afghanistan, Bangladesh, Bhutan, Cambodia, Indonesia, Myanmar, Pakistan, and Uzbekistan, found AI primarily assists in preliminary and final editorial stages. Human oversight remains critical for core journalistic tasks, especially in environments where mistakes can lead to censorship or operational shutdowns. Over 60 percent of surveyed newsrooms lack cohesive AI strategies, relying on individual staff experimentation.
Asian newsrooms are not embracing AI for transformative change, but for assistive tasks. This measured approach is driven by the severe consequences of AI-generated errors, which can extend beyond reputational damage to state-imposed closures or imprisonment of staff. The Asia Centre report identifies that AI tools are mainly used for ideation, summarization, and headline polishing, leaving core reporting and editorial judgment to humans. A significant challenge is the underperformance of major Large Language Models in low-resource Asian languages like Bahasa Indonesia and Bangla. These models frequently mistranslate and generate inaccurate outputs, requiring extensive human correction. Exiled media also face distinct threats, limiting their capacity for AI experimentation, though tailored automation could offer secure tools for transcription and translation. The institutional deficit is clear, with over 60 percent of newsrooms lacking formal AI guidelines, increasing vulnerability to algorithmic bias and surveillance.
