Facebook Hires Team Behind AI Startup in Battle Against Fake News

The social network is interested in Bloomsbury AI's natural language processing expertise

Facebook is adding to its AI team. - Credit by Dianna McDougall for Adweek
Headshot of Patrick Kulp

Facebook is beefing up its artificial intelligence team as its war on fake news continues.

The social network announced Tuesday that it’s hiring the team behind UK startup Bloomsbury AI to improve its ability to comprehend and learn from ordinary language on its platform. Bloomsbury currently runs a web service called Cape that can understand text and answer questions about its content, though it’s unclear what will become of the company’s products after the deal.

TechCrunch, which first broke news of the deal on Monday, reports that the hires will focus on using their natural language processing skills to build tools that will better suss out the fake news and other undesirable content that plagues the platform. Bloomsbury AI’s chief technology officer, Sebastian Riedel, previously co-founded a company call Factmata aimed at battling online misinformation with machine learning.

A Facebook spokesperson declined to comment on whether content issues will be the new team’s top priority.

“The Bloomsbury team has built a leading expertise in machine reading and understanding unstructured documents in natural language in order to answer any question,” Facebook said in its announcement of the deal. “Their expertise will strengthen Facebook’s efforts in natural language processing research, and help us further understand natural language and its applications.”

TechCrunch also reported that Facebook had been in talks to fully acquire Bloomsbury for between $23 million and $30 million before the two companies settled on a sale of the company’s talent without its products or intellectual property. Bloomsbury had previously raised $1.7 million in two investment rounds since its founding in 2015.

Artificial intelligence experts are in extremely high demand as tech companies fight over a talent pool too small to meet all their newfound needs for the technology. Studies estimate that anywhere between 10,000 and 300,000 people worldwide have the know-how to create true machine-learning systems, depending on the complexity of the task.

The hiring comes as Facebook continues to grapple with how to crack down on the sort of misinformation and hoaxes for which it became infamous during the 2016 presidential election. The company has tried out a range of different tools in the past year and a half, including crowdsourced content vetting, machine-learning tools and independent fact-checking protocols.


@patrickkulp patrick.kulp@adweek.com Patrick Kulp is an emerging tech reporter at Adweek.