In a letter Monday to Zuckerberg, the institute noted that Facebook’s terms of service bar the automated collection of public information, a practice that researchers call “scraping,” and the creation of temporary research accounts.
Automated collection allows journalists and researchers to generate statistical insights into patterns and information flows on Facebook’s platform, said Ramya Krishnan, legal fellow at the Knight institute. Sometimes journalists and researchers have attempted to set up temporary research accounts, using a variety of names and biographical attributes, to enable them to assess how the platform responds to different profiles, she said.
But such practices are barred by Facebook’s terms of service, which require that a user “provide accurate information about yourself” and create only one account.
Being able to conduct such research is important because “Facebook’s algorithms are opaque and they have enormous influence on our public discourse,” Krishnan said. “The only way to test how they work is by using tools like temporary research accounts to test algorithms on their platform.”
Journalists who use such techniques may have their accounts suspended and could even risk federal prosecution, said Jameel Jaffer, the Knight institute’s executive director. Both Facebook and the Justice Department have at times interpreted the Computer Fraud and Abuse Act to prohibit violations of a website’s terms of service, he said.
“We are unaware of any case in which Facebook has brought legal action against a journalist or researcher” for violating its terms of service, the letter said. However, in multiple cases, it said, the company has instructed journalists or researchers to stop investigative projects, claiming that they violate Facebook’s user rules.
“The mere possibility of legal action has a significant chilling effect,” said the letter, which was written on behalf of a group of journalists and researchers who have worked for the New York Times, PBS NewsHour, Gizmodo Media Group and Princeton University, and the University of Michigan’s School of Information.
The Knight institute has proposed a “safe harbor” to allow reporters and researchers to undertake investigations in the public interest that are currently barred by the company’s rules.
The revised rules the institute proposes would state that it is not a violation of the firm’s terms of service to collect public information “through automated means” or “by creating or using temporary research accounts” as part of a newsgathering or research project, as long as a number of conditions are met. For instance, the project should be about a matter of public concern, steps should be taken to prevent data theft, the data should not be sold or transferred to any ad network or data broker, and no data should be disclosed that identifies a Facebook user without that user’s consent.
A Facebook spokesman said the company was reviewing the recommendation. In the meantime, Facebook’s head of global news partnerships, Campbell Brown, said in a statement:
“We appreciate the Knight Institute’s recommendations. Journalists and researchers play a critical role in helping people better understand companies and their products — as well as holding us accountable when we get things wrong. We do have strict limits in place on how third parties can use people’s information, and we recognize that these sometimes get in the way of this work. We offer tools for journalists that protect people’s privacy, including CrowdTangle, which helps measure the performance of content on social media, and a new API [software program] we’re launching to specifically analyze political advertising on Facebook.”
Cameron Hickey, a researcher who worked for Miles O’Brien, a science correspondent for the PBS NewsHour, said that after the 2016 election, he decided to investigate the spread of misinformation on social media. He created software to more easily capture data on hundreds of Facebook pages by identifying who clicked on the “like” button on those pages.
“Once I identified this technique, I became aware that scraping content from Facebook would be a violation of their terms of service,” Hickey said. He consulted with NewsHour leadership, who concluded that the legal risks were too great, so he resorted to manual techniques, which were much slower and yielded less data, he said.
Aviv Ovadya, a misinformation researcher who worked for the Center for Social Media Responsibility at the University of Michigan, said the proposed change in the terms of service could actually be beneficial to Facebook by “helping strengthen their platform and helping them identify unintended consequences” they might otherwise miss.
Hickey and Ovadya are among the journalists and researchers whose work prompted the letter.