Newsrooms are increasingly experimenting with AI-based automation to produce sports reports, financial updates and other data-driven stories.
News agencies, including AP, Reuters and AFP, produce thousands of automated stories a year, reinforcing their commitment to delivering timely and accurate information and freeing up resources for other coverage. But how can automation work for local media? And what does it take to introduce it to a newsroom in Latin America where the use of AI in the media is considered rare?
Since March 2022, the 17-year-old local Argentinian newspaper Diario Huarpe (which covers the province of San Juan, with a population of 738,000) has been reporting football stories using automation technology. Weather reports followed in April. Automation has allowed this small newsroom – there is only one reporter covering sports each weekend – to offer more comprehensive coverage of this vast region and coverage of events in neighboring Chile.
“Robots allow us to cover more and give journalists more time and resources for other situations,” says Pablo Pechuan, special projects manager at Diario Huarpe. “For football, if you have the results and the structured data, the robots can write it for you. Then journalists can focus on an interview with the coach, analyzing positions or reporting on the violence around the stadium.
How it works
The automation deployed by Diario Huarpe is managed by technology company United Robots, which first developed it as a way to automate the writing of match reports based on sports data in 2015. It now uses the AI and Natural Language Generation (NLG) to produce automated test reports. , images, charts and maps for its newsroom clients on a range of topics. AI is used to parse structured data from multiple sources and introduce a set of parameters for the output, such as article length and whether to include a title and place first.
“Everything is based on structured data,” says Cecilia Campbell, CMO at United Robots. “It’s not machine learning, which is a likely method and gets increasingly correct with a large enough body of language to base the model on. [With machine learning] you can never guarantee that the facts are in there.
The robot is trained to understand which data points to include in a story. With the correct data, it can produce reports from different angles using the same data points, such as a match report from the home team’s perspective or from the away team’s perspective. While the robot is trained to detect repetitions and if something has already been mentioned, the editor can also provide details of their house style for the robot’s know-how reports to be written.
The NLG tool presents segments of language to robo-reporters — a “sophisticated system of templates,” says Campbell — all originally written by humans. “Quality is the most important thing,” she adds. “We have so many language segments that you can have multiple versions of a text and you’re not going to have identical texts all the time.”
The final output is delivered directly to the publisher’s CMS and Pechuan says minimal, if any, editing is now required before publishing. Diario Huarpe publishes around 250 automated football articles a month, often short action-oriented match reports, with coverage including top-flight football in Argentina and the Copa Libertadores. By comparison, a reporter can produce between four and six stories a day, which greatly increases scale and coverage, says Pechuan. For weather, this increases to around 3,000 articles per month, allowing the newsroom to provide multiple daily updates for certain locations and forecasts.
“We covered these football teams before, but it was difficult because we only had one person for sports on Saturdays and Sundays,” says Pechuan. “Having robots doesn’t mean we don’t need to hire journalists, it’s the opposite. We need more journalists to cover volleyball, hockey and more leagues.
“There normally aren’t the resources or the time to do that in the local media,” adds Campbell of United Robots. “They have many communities they are supposed to serve, but they don’t have enough reporters to cover everything.”
Rather than training the AI to write and understand different languages, the human-led NLG element of the automation process helps perfect the language in reports. A “language tree” is written by an expert in Spanish, for example, for the robot to use based on a set of conditions and rules. It also ensures that Google will not penalize publishers: the quality of language in automated reports must be high.
After the first tests with Diario Huarpe, some adaptation was necessary and there was a “learning period”, explains Pechuan: “Spanish is different in our region. The first versions of the texts were like Madrid Spanish. Now the language is really natural for us.
Early sample match reports were shared with the newsroom and compared side-by-side in a spreadsheet viewed by the tech company and journalists, who could make corrections or suggest language improvements.
It was necessary to introduce a differentiation between the locations. For example, there is a San Francisco in the province of San Juan, but also one in another place, Córdoba, and another in the United States. It was important to clarify these differences both for SEO purposes and to ensure that the public received the correct information, especially for weather updates.
The quality and availability of structured data is crucial for any content that Diario Huarpe can automate.
“Unfortunately, with football and sport in general, there is not a lot of structured data [in Latin America] so we can only cover local teams participating in national leagues,” says Pechuan, who is also trying to find data for youth and women’s leagues. “Data availability is a problem, but we don’t have that problem with the weather. With the weather, you have everything.
A Finnish company provides United Robots with weather data, which is cheaper than sports data. The company is trying to strike deals with data providers, as this can be a challenge for newsrooms, especially smaller ones.
“Data is where it all starts, so it sort of dictates product development,” says Campbell, adding that the availability of structured data sources for different reporting topics in Latin America is currently a barrier to expanding into this market.
The biggest challenge of bringing automation to the newsroom is usually the acceptance of fellow journalists, says Campbell: “A lot of different stakeholders internally have to be on board. They do not replace an old product; they introduce something that did not exist before.
“There are few industries where robots are welcomed by the people who work there. We work with someone who has to get the writing on board. However, journalists used structured data and automated analytics for deeper work. “The combination of reporters and bots is where you get the most value,” she says.
Pechuan acts as a bridge between the newsroom, the product and the technology. He says there was little resistance from journalists in the newsroom, especially once the language was naturalized in the region. Instead, the popularity of some automated reporting with the public has encouraged the newsroom to find better ways to promote and distribute journalists’ original stories.
The future of automation
The automation allows Diario Huarpe to publish weather reports for more than 30 locations across the province, giving its audience more localized and accurate information than before or via weather reports on their smartphones. The homepages of different locations now contain a localized weather report which has brought new search audiences to the sites.
This search interest has led to new landing pages and content designed for audiences interested in particular locations. The newsroom is testing a newsletter based on a more localized offering for 19 different local departments. Growing that audience and developing new product ideas for them balances data costs and keeps technology working, says Pechuan.
The newsroom also uses a third-party AI-powered tool to manage and optimize its social media posting. Without a dedicated social media editor, that’s another boost for efficiency and eliminates duplication of effort, says Pechuan. United Robots has produced automated real estate and business reports for other newsrooms outside of Latin America and Pechuan may consider using automation in other ways, for currency conversion rates given the proximity of the newsroom to Chile or for lottery results, for example. But other automation apps depend on good data being available and needing to meet the needs of local audiences, he says: “There’s not the same demand here in terms of real estate reporting and although there is a solution for automated traffic reports, traffic is not a problem in San Juan.
Oliver is a freelance journalist based in the UK. She has written for the Guardian, BBC, The Week and more. She is a guest lecturer in online journalism at the City, University of London, and works as an audience strategy consultant for newsrooms.