MIT researchers have created an AI-driven cyber security platform that can predict cyber attacks without the risk of human error. Called the AI2, this system combs through data and detects suspicious activity by grouping data into meaningful patterns. When suspicious activity is found, the data is sent to human analysts. In 85% of the cases AI2 made accurate predictions of cyber attacks, assisting human security experts by working through billions of lines of data. Head research scientist, Kalyan Veeramachaneni, hopes this system will become a next line of defence against cyber fraud. “The more attacks the system detects, the more analyst feedback it receives, which, in turn, improves the accuracy of future predictions,” Veeramachaneni says. “That human-machine interaction creates a beautiful, cascading effect.”

Watch the video above for more about AI2.

Source: MIT