Skynet booted as Pentagon straps “Mayhem” autonomous AI into critical systems

After seeing the raw power, and potential, of Mayhem, a new breed of Artificial Intelligence (AI) cyber security “supercomputer” and opne of the world's first Robo-Hackers, and arguably the future of cybersecurity, that can automatically, and it’s easy to argue, autonomously, find and fix vulnerabilities in its own software, and then go off, exploit, and attack an opponent, at DARPA’s Cyber Grand Challenge (CGC), a capture the flag challenge, last year, the Pentagon has finally buckled and officially bought the revolutionary new technology from the small Pittsburg company that made it, ForAllSecure.

The Pentagon, which announced the seven figure purchase earlier this week, will use the cutting edge tech to help secure America’s military systems, including US Navy ships and aircraft and the program will be overseen by the Pentagon’s “start up office,” the Defense Innovation Unit Experimental (DIUx).

The two year contract is part of a program dubbed “Voltron,” which will offer the technology to a variety of different defense agencies in an effort to find code flaws in the operating systems and custom programs used by the US military.

Voltron, which is a multi-vendor effort, that includes, but is not limited to the aforementioned deal, is a new program with the mission of leveraging breakthrough AI to discover issues in military software, and while ordinarily this type of intensive security research would typically need a team of specialised experts, Voltron, as was witnessed at the CGC challenge will be able to complete the task millions of times faster than its equivalent human counterparts.

David Brumley, CEO of ForAllSecure, said the DOD is largely interested in Mayhem for its “self-healing” capabilities, adding that the technology is “extremely scalable” because it’s based on LinuxX86, a popular US Government operating system.

Mayhem’s advantage lies in its ability to create, test and apply patches in real time onto unique systems that are critically important and cannot crash.

“The Defense Department understands the value of autonomy and they have for a long time,” Brumley said, “last year, when they saw the CGC, I think that was really the first time they saw autonomy in cyber … the reception since then has been incredible.”

Although Mayhem also has the potential to be used for offensive hacking purposes, the contract, we’re told, as it’s currently written only outlines specific defensive use cases, including real time and continuous penetration testing and custom patching. A “patch” can be described as a specialty software update that is used to cover holes or flaws in software, which could be leveraged by a hacker.

Brumley said that a handful of foreign military organisations have also reached out to his company since the competition in order to buy and partner with them but that they’d declined their advances, despite currently supplying the new technology to a number of prominent US based private sector technology firms.

As the product is rolled out to the Pentagon in the coming months, ForAllSecure will initially conduct some training, but the hope is that the system can eventually be operated by individuals without cybersecurity training, and ultimately, the purpose of the contract is to provide the US government with a tool that is both highly effective, and low cost.

Anyway, is it just me or should we be wigged out by the fact that the Pentagon has just strapped an autonomous AI called “Mayhem” into its most critical systems? That said though, and perhaps much more importantly, at least we can now all stop talking about the likelihood of Skynet taking over the US Military because let’s face it that won’t happen, it’ll be Voltron… all hail Voltron and its kin.

Elon, hold the rocket, ticket to Mars anyone?