This week's Question: Lars Eidnes, a Norwegian developer, recently created software that uses Recurrent Neural Networks (RNN)—a form of “deep learning”—to write new "clickbait” headlines. After training the software with several million articles from BuzzFeed, Gawker, Jezebel, the Huffington Post, and Upworthy, Eidnes created an entire auto-generated news site called Click-o-Tron, which paired the headlines with photos and short articles, also assembled by the neural network. In early 2015, The Verge reported that the Associated Press has an automated system producing around 3,000 stories per quarter. The Los Angeles Times similarly uses its “Quakebot” algorithm to write short earthquake reports, and Google is finding headlines from newspapers like the Daily Mail to teach its neural networks to parse language.
What do you think? Is robo-journalism valuable?