For once, Facebook and Twitter didn't keep invoking AI as a cure-all
Artificial intelligence got only a few passing mentions, unlike the last few times the tech industry faced Congress.
Roger ChengFormer Executive Editor / Head of News
Roger Cheng (he/him/his) was the executive editor in charge of CNET News, managing everything from daily breaking news to in-depth investigative packages. Prior to this, he was on the telecommunications beat and wrote for Dow Jones Newswires and The Wall Street Journal for nearly a decade and got his start writing and laying out pages at a local paper in Southern California. He's a devoted Trojan alum and thinks sleep is the perfect -- if unattainable -- hobby for a parent.
ExpertiseMobile, 5G, Big Tech, Social MediaCredentials
SABEW Best in Business 2011 Award for Breaking News Coverage, Eddie Award in 2020 for 5G coverage, runner-up National Arts & Entertainment Journalism Award for culture analysis.
In the more than two and a half hours of testimony from Facebook Chief Operating Officer Sheryl Sandberg and Twitter CEO Jack Dorsey (Google was a no-show) before the Senate Intelligence Committee on Wednesday, the term AI specifically mentioned once, with a handful of references to machine learning or bots.
Watch this: Senate takes on deep fakes with Sheryl Sandberg and Jack Dorsey
But Wednesday's session between Dorsey and Sandberg and the senators was far less combative and more collegial. Questions ranged from a look at the value of a customer's data to the tension between authentic and misleading information.
The temptation would be to mention AI, a system that can learn to identify potential fake news and squash at a scale that even an army of humans couldn't match. Zuckerberg in particular drew some criticism for over-using the term AI, which is both vague and highly technical, as a panacea for many problems facing the social networks.
Both executives were more restrained. Dorsey mentioned AI once when discussing how the company is figuring out how to track certain kinds of behavior. Sandberg didn't explicitly say AI, but referenced machine learning a handful of times, specifically to identify fake accounts, as well as deepfakes, or fake videos that look stunningly genuine.
Sen. Jack Reed, a Democrat from Rhode Island, referenced bots when asking about labeling fake accounts. Dorsey said he would consider identifying and labeling certain accounts as "bots," but didn't want to leave the impression that accounts without that label were automatically human.
"You can't have people assume everyone else is not a bot," Dorsey said.
During the second session with the House Committee on Energy and Commerce, Dorsey was specifically asked about the use of AI. He noted that the application of AI at scale is new, and that the company is starting to learn how to use it. He also talked about the need to explain and offer context around some of the decisions of these algorithms.
"We are offloading more and more of these decisions to this technology," he said.
Update, 1:36 p.m. PT: To include additional comments from Dorsey.