A new study conducted by researchers from the University of Cambridge and Stanford University is the first explicit and emphatic demonstration that computers can judge people’s personalities better than the individual can or most people that know the person. Data mining Facebook likes presented a better picture of a person’s personality than the person’s family or friends could.
Only spouses showed a better understanding of a given individual’s personality than a computer algorithm that accessed a person’s levels of openness, conscientiousness, extraversion, agreeableness, and neuroticism. A test population of 17,622 participants was judged by one friend or family member and a sample of 14,410 participants was judged by two friends or family members. The results of the people’s evaluations were compared to the computer analysis and the standard testing of each person. Work colleagues were the worst judges of a person’s personality and were 29 percent less accurate than the computer program.
The study is the first demonstration that artificial intelligence has the capacity to judge humans more accurately than other people if and only if the amount of data is very large. Facebook likes average about 227 per person that uses Facebook. About 500 million people use Facebook regularly enough for the computer personality assessment results to be valid.
The study raises several ethical questions. Will Facebook sell the personality data that it can collect? The data can be used to assess the suitability of a person for a job, a specific training regimen in college, or as a potential mate. The question is, are people ready to give up that much individual choice to a machine even when the machine has proven itself to be a better judge of human personality than people are.
