Research from Cambridge’s Psychometrics Centre, in collaboration with Microsoft Research Cambridge and published in the journal PNAS, found that public traces left by online behaviour on social networking platforms can disclose sensitive and valuable information about users.
A dataset of over 58,000 Facebook users in the US was gathered, the group volunteered their ‘Likes’ and demographic information and underwent psychometric testing to highlight personality traits.
The Facebook ‘Like’ data was fed into algorithms and matched up with information taken from their profiles and the personality tests to infer highly accurate information. The models were proved 88 per cent accurate for determining male sexuality, 95 per cent accurate distinguishing African-American from Caucasian American and 85 per cent accurate differentiating Republican from Democrat.
In addition, Christians and Muslims were correctly classified in 82 per cent of cases, and good prediction accuracy, between 65 and 73 per cent, was achieved for relationship status and substance abuse. The algorithms created more incisive personal profiles through the aggregation of likes across popular culture such as music and television.
The researchers warned of the potential threats to users’ privacy, as such levels of digital exposure could see criminals using predicting software to accurately infer highly sensitive information from such digital traces.
Michal Kosinski, Operations Director at the Psychometric Centre, reckons similar predictions with “remarkable accuracy” as to that gathered from the trial could be made from various online data and it could become “increasingly difficult” for people to control the variety of digital traces they leave behind.
He said: “I appreciate automated book recommendations, or Facebook selecting the most relevant stories for my newsfeed. However, I can imagine situations in which the same data and technology is used to predict political views or sexual orientation, posing threats to freedom or even life.
“Just the possibility of this happening could deter people from using digital technologies and diminish trust between individuals and institutions – hampering technological and economic progress. Users need to be provided with transparency and control over their information.”
Nick Besant, of Panoplia Security Services, a company which provides security consultancy services to companies in the UK and Europe, issued a warning that users need to do more to minimise the risks of identity theft online.
“The topical study which is based on algorithms that infer a user’s identity, suggests your ‘Likes’ disclose far more sensitive and therefore valuable information than previously thought,” he said.
“It is crucial users be both aware and proactive in managing their social media presence, and ensure their own information is not used by a dangerous minority intent on getting our data.”