Facebook tried to ‘alter people’s behavior’ with hundreds of unmoderated tests

Facebook Psychological Tests

Facebook’s social experimentation rabbit hole goes much deeper than any of us could have imagined. The Wall Street Journal reports that the controversial 2012 experiment in which the news feeds of nearly 700,000 users were altered to show either positive or negative posts was just one of hundreds that the Facebook Data Science team has conducted over the past several years.

Facebook was quick to respond, attempting to downplay the effects of the experiment, but the proverbial cat was already out of the bag. A former member of the Data Science team tells the Journal that until recently, there was hardly any moderation when it came to conducting psychological tests on Facebook users. Where a typical research team would need to ask for consent before probing individuals, Facebook depended on language within its fluctuating Terms of Service to take advantage of its user base.

“There’s no review process, per se,” said Andrew Ledvina, a Facebook data scientist from February 2012 to July 2013. “Anyone on that team could run a test. They’re always trying to alter peoples’ [sic] behavior.”

During one innocuous test, thousands of Facebook users received messages alerting them that they were being locked out of their accounts because Facebook suspected they were either bots or using fake names. The users had to authenticate their information in order to log back in, but Facebook knew these users were legitimate — it was just a test of the antifraud system.

In another case, one product manager ran a test without telling anyone else at the company. It got to the point where so many tests were being run that the team began to worry it would start choosing the same users for multiple tests.

Facebook has supposedly implemented much stricter guidelines on experimentation since the 2012 experiment, but the company is considering additional changes in the future. Although an internal review board now reviews research proposals, there’s no indication of how much the experiments have been altered since 2012.

“I’m sure some people got very angry somewhere,” said Ledvina. “Internally, you get a little desensitized to it.”

blog comments powered by Disqus