A Spanish startup is promising to revolutionize readers’ access to often unreported news. The unreported news in question, however, is not overlooked disasters or under-reported tragedies in far-flung countries, but minor league sporting events.
David Llorente, co-founder of Narrativa, said was inspired to develop an AI-powered content generation system after he tried fruitlessly to find coverage of minor league soccer games from other countries in his native Spanish.
“There are people interested in these things, in these leagues, in these kind of sports,” he told Motherboard. “The idea was to focus on regional sports. I wanted to write about football, but about Japanese football in Spanish, to cover this niche.”
Sevilla won with a resounding 20 against Athletic in Nervión, where the sum up eight straight wins at home. Gameiro scored the first one for the locals and closed the scoreboard by converting a penalty kick after Kychowiak was fouled. Athletic was unlucky despite controlling ball possession and wasn’t able to finish any of the numerous chances that they had. – Narrativa game summary
Narrativa is part of the booming automatic content generation industry which uses algorithms to convert data sets into narratives. US company Narrative Science has received investment from In-Q-Tel, the US intelligence investment fund. E-commerce companies, including one in Germany called Idealo, are also investing in automatic content generation in an effort to find a quicker, cheaper and more efficient alternative to copywriters. In media, the Associated Press and Reuters are among those automating stock market coverage, while the Washington Post recently used automation to create briefs that were fed into its Olympics live blog.
When it comes to media, Narrativa, like other content generation services, is limited to data-heavy subjects, meaning sports, finance and economics.
Other competitors in the field include Automated Insights, whose Wordsmithsoftware powers the AP’s automatic content generation, while Thomson Reutersreportedly uses in-house software.
Narrativa’s founders met as colleagues in the computer science department of Spain’s Alcalá University and are firm adherents to using machine learning and Artificial Intelligence—in contrast to a programmatic approach—to generate texts.
“Imagine if you want to add new information about the game, you say I’m going to talk about the weather, how the weather influences the result. In our approach we will just add new data and new examples and the system will learn,” he said. “If you have a programmatic approach you have to define the rules, to code every single line and say here we talk about if the weather was bad or good”
Journalists are key to the process, Llorente added, noting that the company counts three reporters among its staff.
“Our approach helps us to have a lot more continuity and the narratives are potentially much better because they are not repetitive.”
Yet while Narrativa uses journalistic expertise to improve its products, not everyone in the media has been welcoming to their pitch, Llorente said, claiming that there is some inertia among the big media companies in Spain in particular in exploring the possibilities of AI-produced content.
Those adopting automatic content generation are quick to defend against accusations that they are cutting jobs. The AP and the Washington Post have emphasized “robot journalism” as a complementary tool for journalists, rather than something that will be replacing reporters in newsrooms.
Using automated systems, media organizations can produce more content and potentially free up journalists to focus on less repetitive work. In January 2015, the APannounced that it was automatically generating more than 3,000 earnings reports each quarter, noting this represents a “tenfold increase” over what AP reporters and editors created previously.
Llorente is confident that despite some lingering skepticism in the media industry, automatic generation will be everywhere within a few years.