Robot writers will provide us with customised information that are only relevant locally
Are you a football enthusiast but your team is too small for main stream media to cover it? Would you like to receive a daily weather forecast update or pollution information about your area? Claude de Loupy, co-founder and CEO of automated writing and semantic analysis start-up Syllabs, based in Paris, France, brings a solution that can exactly do that. Trained as an engineer, this expert in natural language processing and word sense disambiguation, works on something called robot-journalism. The first such attempt, dates back to March 2014, when the Los Angeles Times, published the first news story written by a robot reporting an earthquake, which took place that morning.
De Loupy says that his goal is “to help [the] media develop their content, enrich it and produce more.” The key to his approach is to make use of the robot writing technology to adapt it to our needs. “We transform data into texts,” he stated while at the 2016 Global Editors Network Summit in Vienna earlier this year.
This automatic writing process makes it possible, for example, to write hyperlocal news on the results of a football match or the financial information of a company, according to de Loupy. Based on a data set the user has to provide, “you have a writing engine configured by our linguists, helped by a machine learning system and then data are transformed into text.”
An added value is the possibility to overcome what they call the language barrier. Indeed, “The mean [tool] in human interaction is language,” he explains. “We use language to talk about many different things. Some people prefer graphics or tables, but a lot of people prefer text. We can give them both.”
Finally, de Loupy believes in the possibility to integrate the speech-technology, someone reading the story produced by a robot. He gives a graphic example: “You are driving, and you can have someone talking about your town.”
Interview by Sabine Louët Podcast editing and cover text: Luca Tancredi Barone
Featured image credit: Luiza Puiu, European Forum Alpbach for the Global Editors Network (GEN)
- All good things come to an end - 30 March, 2018
- Ivo Verbeek: cutting the middle man in language editing - 21 March, 2018
- Podcast: How open science could benefit from blockchain - 31 January, 2018