Facebook butts heads with Instagram researchers studying photo site's algorithm

AlgorithmWatch, a German research and advocacy group, shut down its Instagram monitoring project after what it says was a "thinly veiled threat" from Facebook. But the social network says it made no such threat and that the group's project ran afoul of Facebook policies around data collection.

The advocacy group says it's "committed to evaluating and shedding light on ... algorithmic decision-making processes that have social relevance" and that its project found that Instagram prioritizes posts that feature people who are "scantily clad" and that politicians' posts were seen by more people when those posts showed a politico's face instead of text. 

In a blog post Friday, the researchers said they shut down the Instagram project on July 13, after a May meeting with Facebook, which owns Instagram. At that meeting, they said, Facebook told AlgorithmWatch it had violated Facebook's terms of service, which prohibit the automated collection of data. According to the group, Facebook said it would "mov[e] to more formal engagement" if the issue wasn't resolved, which the researchers took as a threat of legal action. 

Facebook says it didn't threaten any legal action against AlgorithmWatch and wanted to work with the organization to find a way to continue the research. 

"We had concerns with their practices," a Facebook spokesperson said in an email Friday, "which is why we contacted them multiple times so they could come into compliance with our terms and continue their research, as we routinely do with other research groups when we identify similar concerns."

As part of the Instagram project, AlgorithmWatch developed an add-on that scraped volunteers' Instagram newsfeeds to study how the social network "prioritizes pictures and videos in a user's timeline." The researchers contend that the add-on's users volunteered their feed data to the project and that since the project's launch, in March 2020, about 1,500 volunteers had installed the add-on.

Earlier this month, Facebook disabled a similar research project at New York University, saying it violated the social network's terms around data gathering. The NYU Ad Observatory used an add-on to collect data regarding what political ads were shown in a user's Facebook feed.

News about the shutdown of AlgorithmWatch comes as there's been intense scrutiny on social networks, the misinformation found on them and the effect they have on individuals and society.

For its part, Facebook has had to be careful with how it manages the data of its users, particularly following 2018's Cambridge Analytica scandal, in which an outside firm harvested information from 50 million Facebook accounts without their permission. That scandal led to Facebook CEO Mark Zuckerberg being called before Congress to testify about the social network's data privacy policies. And it played a part in Facebook agreeing, in 2019, to pay a $5 billion fine to the US Federal Trade Commission over privacy violations. Under that settlement, Facebook must certify that it's taking steps to protect user privacy.

The Facebook spokesperson said Friday that the company makes it a point to cooperate with researchers. "We collaborate with hundreds of research groups to enable the study of important topics, including by providing data sets and access to APIs, and recently published information explaining how our systems work and why you see what you see on our platform."

AlgorithmWatch, on the other hand, accused Facebook of "weaponizing" its terms of service. "Given that Facebook's terms of service can be updated at their discretion (with 30 days' notice), the company could forbid any ongoing analysis that aims at increasing transparency, simply by changing its terms," the group said in its blog post.