Journalism and artificial intelligence: editorial trust as a limit to automation

AI can summarize documents, assist in searches, sort information, suggest headlines, transcribe interviews, detect patterns in large volumes of data, and speed up tasks that previously required hours of manual work. But that promise of efficiency also opens a more delicate discussion for the media: what happens when automation begins to intervene in processes where verification, editorial judgment, data protection and the relationship of trust with audiences matter.

That was one of the axes that crossed the presentation of the Mexican Juan Carlos López CalvetDirector of Data & AI at Schibsted News Media, one of the largest Scandinavian media groups. The specialist participated in a new edition of the cycle in Buenos Aires Innovation, Technology and Journalismorganized by Personal within the framework of the Journalist’s Day.

During the meeting, which brought together journalists and media representatives from all over the country, López Calvet spoke about the pragmatic, responsible use of artificial intelligence aligned with rigorous editorial standards. He also demonstrated practical uses to apply AI in journalistic tasks and routines.

A contribution by no means minor in an increasingly turbulent ecosystem: that of the news, crossed by the fake news and the excess of unchecked publications.

The central point of the exhibition was how to bring AI to a newsroom without reducing it to a tool to produce more in less time. López Calvet highlighted four vertical that guide the use of these technologies in the Norwegian group: “AI to work better, AI to serve the audience better, AI to create new products and AI to responsible use as a scale condition.”

That last aspect works as a warning for an industry that often adopts technology through the door of productivity. In media, automation can improve workflows, reduce repetitive tasks, and open up new ways to distribute information. But of course: It can also amplify errors, biases, opacity, or poorly designed editorial decisions.

For López Calvet, the value of AI in a newsroom should be measured by its contribution to the quality, understanding and scope of journalism. The logic of saving time or reducing costs is incomplete when the final product affects the credibility of the medium or the public’s ability to better understand a topic.

Along these lines, the specialist highlighted that the incorporation of AI requires usage frameworks, limits for sensitive content, metrics aligned with public value and humans with decision-making power throughout the process. The editorial governance It thus appears as a condition for scaling automated tools within a journalistic organization.

López Calvet also pointed out three tensions that every editorial office must face in these new work scenarios: the speed of the processes versus data security, autonomy versus trust in traditional infrastructure, and the risk of experimentation versus editorial trust.

The first tension is one of the most visible. AI tools allow us to advance faster, although they often work with sensitive information: sources, internal documents, databases, embargoed materials or content not yet published. In a newsroom, the use of these systems requires clear criteria about what can be uploaded, where the information is processed, and what controls exist over the results.

The second tension appears when the teams They seek autonomy to experiment. AI allows journalists, editors, designers or developers to test their own solutions without always depending on large technological projects. That flexibility can accelerate innovation, but it also requires reliable infrastructure, traceability and oversight.

The third tension touches the heart of the craft. Experimenting with AI can open up new products, formats and services for audiences. At the same time, any errors in sensitive issues can damage editorial trust, one of the most difficult assets for a medium to recover.

López Calvet closed his presentation with an idea that summarizes his approach: “Artificial intelligence creates value when the chain improves of tasks in their entirety, not on an isolated task.

The phrase points to a key difference, which is that, in a newsroom, automating part of the process can generate local efficiency and at the same time problems in the final result. A summary can save time, although it requires verification. A personalized recommendation can improve a user’s experience, although it can also fragment the public agenda. A text generation tool can speed up drafts, although editorial judgment is still necessary to prioritize, contextualize and publish.

Given the most recurring question, namely, If certain roles in journalism are going to die, From their perspective, routine and repetitive tasks will disappear, especially those associated with mechanical processes. The journalist, on the other hand, retains a central place in functions that require judgment: verify, edit, interpret, explain, decide what matters and take responsibility for what is published.

That approach also includes the explanatory capacity of the models. For media, understanding why a tool suggests, classifies or prioritizes certain content becomes relevant when those decisions affect the circulation of information or the relationship with audiences.

The opening of the meeting was led by Pedro López Matheu, director of External Communications, Sustainability and Personnel Resources. There he presented a new edition of the IA Journalism guidean initiative of the Redacciones5G program that incorporates recent advances in artificial intelligence in journalistic practice.

“We have the mission, through this cycle and our institutional program Redaciones5Gto accompany the evolution of journalism. Today we present a new edition of the AI ​​Journalism guide that incorporates the advances of artificial intelligence, already integrated into daily life and journalistic practice,” said López Matheu during the presentation.

Marcos Calligaris, Content Director of Cadena 3, and Patricio Caruso, General Digital Editor of TN, also participated in a local reflection panel.

Redacciones5G was born more than ten years ago as a training initiative promoted by Personal to accompany digital transformation processes in media in Argentina and the region. According to the company, the program “toured more than 500 media outlets, trained more than 10,000 journalism and communication professionals, and reached more than 3,000 students in related careers.”

The substantive debate, however, exceeds a specific day or tool. AI is already integrated into part of daily life and is beginning to modify the morphology of newsrooms, professional profiles, work tools and the relationship with audiences.

The question for the media is how to automate without diluting what sustains its legitimacy: trust. This question also affects many other professions and trades affected by the artificial intelligence revolution.