Facebook’s algorithm can tell whether you take drugs or not — even if you don’t post about your extracurricular activities. According to a new study, the social network’s can figure out if you drink or do drugs, based solely on likes and status updates.
The study was conducted by researchers at USC Berkeley who analyzed the activities of 11 million Facebook users, as well as an extra 22 million status updates from 150,000 people. The scientists used a machine-learning software that analyzed certain keywords including “fuck,” “shit,” “hate,” “kill,” “blood,” “pain,” and “sex.” It also tracked likes, finding that people who are into anime are less likely to binge-drink, while if you’re a big V For Vendetta fan you’re more likely to reach for the booze.
The study’s intention was to identify either those at risk or those currently suffering from substance abuse disorder as part of drug prevention research, and it delivered with a crazy 80 to 86 percent accuracy. Though it’s unclear how exactly the study’s findings will be implemented to prevent drug use, it does shine a light on just how much about you can be gleaned from a seemingly innocuous online presence.
In other news, weed could help prevent brain diseases.
- Source: Thump
- Main & Featured Image: Thump