One of the world’s best universities is coming under increased pressure over how it handled concerns raised about the mass data harvesting at the heart of the Facebook and Cambridge Analytica scandal, after a cache of emails revealed heated exchanges between senior academics involved.
The May 2014 emails seen by the Guardian reveal staff at Cambridge University, including a departmental director, challenged the psychologist who collated data belonging to more than 50 million Facebook users over his plans to strike a deal with the parent firm of Cambridge Analytica. In one email confronting Aleksandr Kogan, his activities are branded a “get rich quick” scheme.
The correspondence reveals the deep anxiety among university psychologists over the activities of Kogan, an assistant professor who, at the time, was involved in talks with Strategic Communication Laboratories (SCL), the parent company of Cambridge Analytica.
One of Kogan’s contemporaries raised specific concerns over whether university resources including computers and staff had been used to conduct the controversial tests that yielded the data that ultimately ended up in the hands of Cambridge Analytica.
Cambridge University has said it sought and received reassurances from Kogan that no university data, resources or facilities were used as the basis for his work. Kogan remains employed at the university.
Kogan and his business partner, Joseph Chancellor, who later went on to work for Facebook, where he remains, harvested data using a personality app developed through Kogan’s commercial enterprise Global Science Research (GSR).
The emails from late May 2014 show Kogan initially tried and failed to elicit the support of two other academics at the university, who had amassed their own trove of Facebook data from volunteers by hosting personality apps on the platform. This did not include data from friends.
Prof John Rust, the director at Cambridge University’s Psychometrics Centre, which had pioneered the study of psychology through large-scale data analysis, wrote to Kogan after a face-to-face meeting about a dispute with the two academics.
Rust accused Kogan of trying to make $1m in “personal profit in terms of asset and data” from the scheme, while only reimbursing his fellow psychologists, Dr Michal Kosinski and Dr David Stillwell, who had led much of the cutting-edge research, with $100,000.
“When they queried this, you replied that, unless they agreed, you would go on to develop the project without them as the ideas were now in the public domain,” Rust told Kogan.
“None of this seems right to me,” he added, pointing out that Kosinski and Stillwell had honed the department’s work on Facebook and personality traits “for the furtherance of academic research into the prediction of personality and behaviour from online footprints”.
It is understood Kosinski and Stillwell planned to channel any funds from involvement in the SCL contract into university research. However, negotiations broke down. It was at that point that Kogan and Chancellor opted to harvest their own data through GSR, raising fears they were exploiting the research of their colleagues.
“I must insist that you do not take advantage of your position in the university to fund a ‘get rich quick’ scheme at our expense,” Rust told Kogan in the email. He noted that Kosinski and Stillwell had made another tranche of data available to Kogan and Chancellor “explicitly for academic research purposes only”.
Rust concluded his email to Kogan: “From our conversation it seems to me that it is unlikely that we are going to agree on this. As well as the relevant line management, I am copying in legal services in the belief that they may be able to assist in independent arbitration before matters get seriously out of hand.”
The director of the Psychometrics Centre also wrote directly to Chancellor, asking for an assurance that any participation of commercial projects “such as the one with SCL Group” would not use data or other resources gleaned by the university.
Replying to Rust, Kogan said: “Your accusations are highly inaccurate, unfounded, and quite frankly, bullying.”
Kogan went on to insist that the university’s resources or intellectual property were not part of the deal with SCL. “The amounts you quote represented middle of negotiation amounts,” he added, saying SCL was unwilling to pay those sums for models they could build in-house.
“We have used no data and no algorithms from Psychometrics Centre. We have built all our models using our own data, collected using our own software,” Kogan added.
Kosinski then told Rust that he had been informed that Kogan was working on the SCL project “during his office hours, using university computers and servers, and the work of his interns and post-docs”.
“This is a clear violation of the assurances both Dr Kogan and Dr Chancellor gave to the university,” wrote Kosinski, the deputy director of the Psychometrics Centre.
“The continuation of his unethical behaviour is really disappointing. It shows that the assurances of the faculty member of the psychology department cannot be trusted and the department does not take any steps to solve this issue.”
Kosinski added in the note that while the concerns the academics had flagged over Kogan “were really serious” and the damage caused by his action “very significant” their complaint appeared to have been “effectively dropped” by the department.
“I believe that this situation is really disturbing to the culture of our department and destroys the good name of the university,” Kosinski added.
Kogan has offered an explanation of his actions in an interview with the BBC and an email to his Cambridge colleagues seen by the Guardian. Kogan told BBC Radio 4 that he thought he was being “basically used as a scapegoat by both Facebook and Cambridge Analytica”.
Speaking to the Guardian, Kogan said: “When we initially envisioned doing the project with SCL, we had planned for it to be collaboration with Michal Kosinski and David Stillwell at the Psychometrics Centre.
“Even though I very much wanted to work with Michal and David, SCL felt that their position of a fee of $500,000 for the modelling portion of the project was not warranted. And so SCL asked me to remove Michal and David from the project.
“I did my best to find a middle road, but neither SCL nor Michal [or] David ended up moving to a middle ground.
“John Rust, the director of the Psychometrics Centre, then made a complaint to the university on behalf of Michal and David … and even though the university’s investigation resulted in no support for John’s claims, I still lost valuable collaborators that I enjoyed working with.
“In the end, I believe the Psychometrics Centre wanted to get back on the project, even though SCL had moved on based on their negative experience with them.”
Cambridge Analytica used the data from the profiles harvested by Kogan to build a powerful software program to predict and influence voters’ choices at the ballot box.
A whistleblower revealed in the Observer on Sunday how Cambridge Analytica used the personal information – taken in early 2014 – to build a system that could profile individual US voters, in order to target them with personalised political advertisements.
The university also revealed that in 2015 Kogan applied for ethical approval to use data collected by his commercial enterprise for his academic research but this was rejected.
Cambridge University said: “We can confirm that there was a dispute between Drs Kosinski [and] Stillwell and Dr Kogan over the question of intellectual property. As a result, we commissioned advice from external lawyers to review the scope of IP rights in the software app ‘myPersonality’. However, as this is legally privileged, we are unable to provide further details.
“We understand that Dr Kogan subsequently made no further use of data from myPersonality. Dr Kogan signed a legal assurance that no data, datasets, models, or software created or developed by doctors Stillwell and/or Kosinski or by any other students or staff at the university, or derived from them, had been or would be used by or for the benefit of SCL.”
