No security solution is perfect. Each one has its own set of pros and cons. For example, relying completely on an automated solution is thorough, but it will flag plenty of threats that aren’t really threats (aka, false positives). Meanwhile, a human overseeing security is great for spotting worrisome trends, but a human can’t possibly catch every single attack. With this dynamic in mind, a team of researchers from MIT has successfully blended the two.
The team, hailing from MIT’s Computer Science and Artificial Intelligence Laboratory, developed a security platform combining the strength of human involvement with the effectiveness of artificial intelligence. Given the name A.I.Â², this new platform is capable of detecting 85 percent of incoming attacks while reducing reported false positives by 20 percent.
How A.I.Â² combines AI and human involvement is actually quite genius. To begin, the “untrained” machine reports a sample set of results to a human expert. These results are informed by machine learning and are checked against the 200 most important problems. The human expert then checks over and corrects the machine’s work, with the intention of finding mistakes and false positives. Subsequently, the human-checked results are inputted to A.I.Â², the machine learns, and the next set of tests will have fewer errors. This process is then repeated with the security platform adapting each time to look for what its human teacher looks for.
In a paper presented at the IEEE International Conference on Big Data Security, research scientist Kalyan Veeramachianeni summarized, “You can think about the system as a virtual analyst. It continuously generates new models that it can refine in as little as a few hours, meaning it can improve its detection rates significantly and rapidly.”
Of course, the precedent of effectiveness set here by combining human expertise and oversight with A.I. has far-reaching ramifications for other A.I. initiatives currently in development; initiatives that have both humanity’s best and worst interests in mind. Although, for now, those of us in the IT world are celebrating this A.I.-enabled technology for being one giant leap forward in security analysis accuracy.
It’s important to keep in mind that a platform like A.I.Â² is still years away from being available to the average business. Therefore, we recommend that you stay on top of your network security by following best practices, which combines automated software solutions and human vigilance in order to combat the latest threats. Think Tank NTG is standing by to assist you with all of your company’s network security needs. To equip your business with the most comprehensive IT security solutions on the market, call us today at 800-501-DATA.