The Boston Globe is looking for new ways to infuse artificial intelligence into how it sources and distributes stories through a new partnership project with Brooklyn startup hub Newlab.
The two companies debuted a startup today called Applied XLabs that will build tools for media companies—and potentially adjacent information industries—to glean insights from large datasets, produce content on a larger scale and better target readers. Francesco Marconi, former R&D Chief at the Wall Street Journal and head of AI strategy at the Associated Press, will lead the new business as founder and Newlab’s svp of AI & Data.
“AI-based technologies are changing the way information is sourced and, more fundamentally, the way we communicate,” Marconi said in a statement. “We have a responsibility to advance machine learning and natural language processing tools for the benefit of society.”
Housed in an 84,000-square-foot expanse of the Brooklyn Navy Yard, Newlab counts around 150 startups as members, providing them access to expensive AI and robotics tools. The portfolio is heavy on AI-related ventures, but it also encompasses startups based around other emerging fields, like blockchain, extended reality and biotech.
The partnership comes as the Globe has been expanding the geographical breadth of its coverage into areas like Rhode Island looking to fill gaps left by the decline of local papers. The 148-year-old Globe believes that new analytics tools will help guide decisions about how to best cater that coverage to expanded audiences.
The paper isn’t the only media outlet exploring how AI might make newsgathering and content distribution more efficient. The New York Times’ data unit has rolled out a variety of machine learning tools, and The Washington Post has been steadily weaving its own homegrown AI into everything from targeting native ad content to writing basic stories.
While AI or automated programs might be able to produce rote, template-friendly articles like earnings reports and sports scores, the technology isn’t yet practical for any more complex stories. That’s changing fast, however, as breakthroughs in natural language processing research have enabled machine learning systems to spit out realistic-sounding copy—though it’s not quite so adept at sticking to any real facts.