Argentine editors embrace AI tools but retain human editorial control

Editors from LA NACION, Clarín and Infobae said artificial intelligence should be used to support reporting and newsroom operations, while journalists retain authority over editorial choices and publication. The shared position emerged during “Fear of change versus reinvention: AI in Argentine newsrooms,” a first-day panel at Media Party Buenos Aires 2026, held at Ciudad Cultural Konex.

Gastón Roitberg of LA NACION, Javier Kraviez of Clarín and Opy Morales of Infobae appeared on the panel, moderated by Guadalupe López of Chequeado. They agreed that editorial decisions should not be delegated to machines, even as the technology is embedded more deeply in newsroom work.

LA NACION reported that Clarín had run a months-long training programme, with an external consultancy and its Big Data team, to address journalists’ concerns about AI and job security. LA NACION, meanwhile, chose a cross-functional team rather than a separate AI department and published an internal guide that other media organisations can adapt.

Roitberg said the initial period of experimentation had passed. He argued that newsrooms now needed procedures to supervise and audit automated processes, enabling professionals to spot errors and control results. He also urged organisations to begin innovation work by defining the problem to be solved rather than choosing technology first.

Infobae has taken a platform-led approach through Scribnews, its proprietary system integrated with the company’s content-management system. The platform, launched in 2023, supports drafting, summaries, headline generation, search optimisation, transcription, translation, editing and archive-image searches, but requires editorial review before publication.

At the panel, Morales said an AI model could contain data and information without possessing journalistic sources. Infobae has also replaced static PDF training manuals with an AI assistant that answers questions about platform functions and incorporates recordings of training sessions as its knowledge base expands. The company limits AI-generated visuals and does not use AI to recreate current events or breaking news, instead dispatching reporters and cameras.

Kraviez warned that lower production costs alone would not ensure a publisher’s sustainability, arguing that readers pay for relevant journalism from trusted organisations. He also cautioned that algorithms could come to define the public agenda.

The debate reflects wider use of the technology among Argentine publishers. An Adepa study based on 20 interviews with media professionals found that 29% of participating outlets used AI for writing or editing, headlines or translation; 38% used it for transcription; and 40% for audio versions or article summaries. Adepa recommended human oversight, intellectual-property protection, audience transparency and continuing training.

Audience exposure is also widespread. The Reuters Institute found that 67% of respondents in Argentina had noticed at least one AI feature on a news website or app in the preceding week. Its cross-country survey found that 9% of respondents used generative AI to get news in 2026, up from 6% in 2025 and 3% in 2024.