In Mendoza, Argentina, the digital news outlet Diario UNO is deploying an artificial intelligence tool, internally named “Tuki,” to automate the drafting of news articles from audio recordings and written documents. The initiative, born from a need to address both unstructured AI usage within the newsroom and the burden of repetitive tasks on experienced journalists, represents a growing trend across Latin American media organizations seeking practical applications for AI.
The project began as a prototype converting audio from Radio Nihuil into initial drafts, but quickly expanded to become a resource accessible to journalists throughout the group. This transition, however, presented organizational challenges, requiring coordination between editorial and technical teams and a shift in perspective – viewing implementation as a cultural change rather than a purely technological one, according to information released by WAN-IFRA.
Today, Tuki generates draft articles while adhering to the outlet’s established style guide and editorial standards. Crucially, the team adopted a “human in the loop” approach, ensuring that automation serves as an efficiency layer, with journalistic judgment and human editing remaining central to the process. This approach reflects a broader concern within the industry: integrating AI without sacrificing editorial focus or audience trust, as highlighted by WAN-IFRA’s work with newsrooms across the region.
The development of Tuki underscores a key finding in the exploration of AI in Latin American journalism – the most significant hurdle isn’t the technology itself, but rather its effective integration into existing workflows. News organizations are prioritizing concrete needs and practical learning, moving away from generic solutions and toward strategies that align technology with their journalistic missions. WAN-IFRA has facilitated this process through initiatives like the LATAM Newsroom AI Catalyst, Cohort 2, supported by OpenAI, which aims to accompany media organizations in the practical exploration of artificial intelligence.
The primary lesson learned from the Tuki project, according to Diario UNO, was the importance of systematization. AI transitioned from being a fragmented, individual practice to a shared process governed by clear rules and objectives. The next phase involves scaling the Tuki platform to serve as a broader editorial support system, further streamlining workflows.
This pragmatic approach to AI adoption is characteristic of the Latin American news ecosystem, which emphasizes experimentation, strategic standardization, and accountability in both the tools and their developers, as noted in a recent report by Aspen Digital. The focus on audience needs and impact is also evident in the work of organizations like Data_Labe in Brazil, which uses data journalism to investigate social issues and demystify AI for marginalized communities.
The increasing adoption of AI in Latin American journalism coincides with a broader regional discussion about its implications, as evidenced by the Montevideo Declaration on Artificial Intelligence and its Impact on Latin America, disseminated at the Latin American Meeting in Artificial Intelligence Khipu in March 2023. The impact of AI on journalism is no longer in question, with more and more newsrooms actively adopting initiatives to utilize and manage it, according to reporting from the LatAm Journalism Review.
Diario UNO has not yet announced a firm timeline for the full-scale rollout of the expanded Tuki platform.
