Volume of data suggests trusted partnership with Aleksandr Kogan, says analyst
Before Facebook suspended Aleksandr Kogan from its platform for the data harvesting “scam” at the centre of the unfolding Cambridge Analytica scandal, the social media company enjoyed a close enough relationship with the researcher that it provided him with an anonymised, aggregate dataset of 57bn Facebook friendships.
Facebook provided the dataset of “every friendship formed in 2011 in every country in the world at the national aggregate level” to Kogan’s University of Cambridge laboratory for a study on international friendships published in Personality and Individual Differences in 2015. Two Facebook employees were named as co-authors of the study, alongside researchers from Cambridge, Harvard and the University of California, Berkeley. Kogan was publishing under the name Aleksandr Spectre at the time.
A University of Cambridge press release on the study’s publication noted that the paper was “the first output of ongoing research collaborations between Spectre’s lab in Cambridge and Facebook”. Facebook did not respond to queries about whether any other collaborations occurred.
“The sheer volume of the 57bn friend pairs implies a pre-existing relationship,” said Jonathan Albright, research director at the Tow Center for Digital Journalism at Columbia University. “It’s not common for Facebook to share that kind of data. It suggests a trusted partnership between Aleksandr Kogan/Spectre and Facebook.”
Facebook downplayed the significance of the dataset, which it said was shared with Kogan in 2013. “The data that was shared was literally numbers – numbers of how many friendships were made between pairs of countries – ie x number of friendships made between the US and UK,” Facebook spokeswoman Christine Chen said by email. “There was no personally identifiable information included in this data.”
Facebook’s relationship with Kogan has since soured.
“We ended our working relationship with Kogan altogether after we learned that he violated Facebook’s terms of service for his unrelated work as a Facebook app developer,” Chen said. Facebook has said that it learned of Kogan’s misuse of the data in December 2015, when the Guardian first reported that the data had been obtained by Cambridge Analytica.
“We started to take steps to end the relationship right after the Guardian report, and after investigation we ended the relationship soon after, in 2016,” Chen said.
On Friday 16 March, in anticipation of the Observer’s reporting that Kogan had improperly harvested and shared the data of more than 50 million Americans, Facebook suspended Kogan from the platform, issued a statement saying that he “lied” to the company, and characterised his activities as “a scam – and a fraud”.
On Tuesday, Facebook went further, saying in a statement: “The entire company is outraged we were deceived.” And on Wednesday, in his first public statement on the scandal, its chief executive, Mark Zuckerberg, called Kogan’s actions a “breach of trust”.
But Facebook has not explained how it came to have such a close relationship with Kogan that it was co-authoring research papers with him, nor why it took until this week – more than two years after the Guardian initially reported on Kogan’s data harvesting activities – for it to inform the users whose personal information was improperly shared.
And Kogan has offered a defence of his actions in an interview with the BBC and an email to his Cambridge colleagues obtained by the Guardian. “My view is that I’m being basically used as a scapegoat by both Facebook and Cambridge Analytica,” Kogan said on Radio 4 on Wednesday.
The data collection that resulted in Kogan’s suspension by Facebook was undertaken by Global Science Research (GSR), a company he founded in May 2014 with another Cambridge researcher, Joseph Chancellor. Chancellor is currently employed by Facebook.
Between June and August of that year, GSR paid approximately 270,000 individuals to use a Facebook questionnaire app that harvested data from their own Facebook profiles, as well as from their friends, resulting in a dataset of more than 50 million users. The data was subsequently given to Cambridge Analytica, in what Facebook has said was a violation of Kogan’s agreement to use the data solely for academic purposes.
In his email to colleagues at Cambridge, Kogan said that he had created the Facebook app in 2013 for academic purposes, and used it for “a number of studies”. After he founded GSR, Kogan wrote, he transferred the app to the company and changed its name, logo, description, and terms and conditions. CNN first reported on the Cambridge email. Kogan did not respond to the Guardian’s request for comment on this article.
“We made clear the app was for commercial use – we never mentioned academic research nor the University of Cambridge,” Kogan wrote. “We clearly stated that the users were granting us the right to use the data in broad scope, including selling and licensing the data. These changes were all made on the Facebook app platform and thus they had full ability to review the nature of the app and raise issues. Facebook at no point raised any concerns at all about any of these changes.”
Kogan is not alone in criticising Facebook’s apparent efforts to place the blame on him.
“In my view, it’s Facebook that did most of the sharing,” said Albright, who questioned why Facebook created a system for third parties to access so much personal information in the first place. That system “was designed to share their users’ data in meaningful ways in exchange for stock value”, he added.
Whistleblower Christopher Wylie told the Observer that Facebook was aware of the volume of data being pulled by Kogan’s app. “Their security protocols were triggered because Kogan’s apps were pulling this enormous amount of data, but apparently Kogan told them it was for academic use,” Wylie said. “So they were like: ‘Fine.’”
In the Cambridge email, Kogan characterised this claim as a “fabrication”, writing: “There was no exchange with Facebook about it, and … we never claimed during the project that it was for academic research. In fact, we did our absolute best not to have the project have any entanglements with the university.”
The collaboration between Kogan and Facebook researchers which resulted in the report published in 2015 also used data harvested by a Facebook app. The study analysed two datasets, the anonymous macro-level national set of 57bn friend pairs provided by Facebook and a smaller dataset collected by the Cambridge academics.
For the smaller dataset, the research team used the same method of paying people to use a Facebook app that harvested data about the individuals and their friends. Facebook was not involved in this part of the study. The study notes that the users signed a consent form about the research and that “no deception was used”.
The paper was published in late August 2015. In September 2015, Chancellor left GSR, according to company records. In November 2015, Chancellor was hired to work at Facebook as a user experience researcher.
Neither Facebook nor Chancellor has responded to numerous queries about his knowledge of Kogan’s and GSR’s activities.