Artificial Intelligence detects 85% of network attacks, becoming smarter

Artificial Intelligence (AI) systems by MIT researchers are able to detect up to 85% of the potential cyber attacks. It is even smarter because of the ability to learn as a human.

A team of scientists from the MIT Institute's artificial intelligence and computer science lab is building a defense system that is immune to security and privacy threats. It is an AI system that can instantly detect up to 85% of the potential network attacks as they occur.

Self-learning model of AI2 system. Photo: The Next Web.

The system is named AI2. They will investigate the abnormal activity of the data and present their findings to the person analyzing and determining what event is actually the attack. This information will be recorded in the AI2 system for subsequent investigations. That's how to teach the computer self-learning of this system.

According to the scientists who are building AI2, the system is scored three times faster than previous systems as well as more than the deviation of authentication. The more data for AI2 analysis, the more accurate the predictions.

According to the team, the AI2 can expand to billions of lines per day to protect large networks.

See more videos on how the AI2 works: