Lessons learned from integrating AI in a local newsroom in Argentina

May 30, 2024 in Media Innovation
Argentine newsroom

Two decades after the internet revolutionized how we access and distribute information, artificial intelligence (AI) has the potential to boost creativity and automation in the newsroom on an unprecedented scale.

As a journalist, professor, and digital media consultant working in the media industry for the last 20 years, I experienced firsthand this century’s profound transformations in news consumption. And for the past three years, I’ve dedicated my work to better understanding how AI will impact journalism. I’ve read research, written articles, taught classes, given talks, developed experimental projects, participated in interviews, and trained and consulted for media outlets and journalists from Argentina and greater Latin America.

In 2023, I worked with Todo Jujuy, a local Argentine media outlet in the province of Jujuy, to help them begin integrating AI into their workflows. This consultancy was part of "Impulso Local," a program that trained 235 media outlets throughout Argentina on various topics, implemented by the Argentine Association of Journalistic Entities (ADEPA), the International Center for Journalists (ICFJ) and Meta

I worked with Todo Jujuy for eight months, from the start seeking to make clear that implementing AI in the newsroom involved not only the introduction of new technologies but also a cultural shift and a reconfiguration of existing workflows. This was evident to those of us who worked in media as it was being digitized in the early 2000s.

In meetings with the Todo Jujuy team, we identified strategic areas that would benefit from AI, selected the appropriate tools and programs for their editorial needs, and trained the staff on new processes.

In this article I document the process developed alongside Todo Jujuy, with the goal that this can be shared as a resource with lessons learned for other local media looking to utilize AI in the newsroom.

Integration first steps

When we began the integration process I asked Todo Jujuy team members to share their experiences and expectations around AI, and write down goals they hoped to achieve. Here are some examples of what they shared:

"The production of articles related to services, such as weather, traffic, sports results, or currency exchange rates, is one of the tasks that our media could automate, thus allowing our team of journalists to focus on stories of relevance to our audience and community."

"In this way, through large-scale production and relevant editorial content, we aim to attract new users, strengthen the bond with current ones, and thus improve our advertising sales revenue."

I then trained the team on AI concepts, and its potential uses and challenges. Together, we also assessed the viability of specific AI technologies, taking into account, for instance, cost, scalability, and how their impact could be maximized in the newsroom. 

Todo Jujuy’s reporters, editors, developers, and content manager all attended and involved themselves in the meetings, demonstrating a cross-organizational commitment that laid the foundation for the newsroom to effectively integrate AI.

We initially identified six editorial and product areas with potential for development with AI. Among them were the automation of sports coverage and publishing in the content management system, a chatbot reading assistant, and audio versions of articles to improve accessibility.

Taking into account that Todo Jujuy runs a radio station and TV channel, as well, I also proposed automatically repackaging radio segments as newsletters, podcasts and articles for the website.

Throughout, we shared and discussed industry reports regarding AI in the newsroom,  learning from how other media outlets were using new technologies. These exchanges better enabled the Todo Jujuy team to make more informed decisions about their own work. 

One of the primary goals we pursued was to move away from reporting news already covered by other outlets, and to focus instead on covering local news and stories that could offer greater impact and value for Todo Jujuy’s audience.

Experimentation and rollout

After identifying the tools and programs to adopt, determining the available financial resources to do so, and selecting the journalists who would be involved, Todo Jujuy began experimenting with automating sports coverage, and formally incorporating AI into other newsroom workflows.

Later, I worked with the editorial team to develop guidelines for using AI in the newsroom. We reviewed reports and statistics on the topic to identify usage patterns that could inform the policies we were developing. The resulting guidelines aimed both to provide an ethical and practical integration framework for journalists, and also to inform how to transparently use the technology for audiences.

The Todo Jujuy team then piloted the automated products, testing their effectiveness and identifying adjustments to be made before general implementation.

The collaboration across departments at Todo Jujuy in creating the guidelines ensured that the process was aligned with the outlet’s needs, enhancing the likelihood that any adoption of new technology would contribute productively and sustainably to innovation in the newsroom.

This article is the first in a two-part series.

Photo created with ChatGPT.