Pentagon Doesn't Need Human Hackers Any More, Develops New AI Tool for Cyber Warfare

In a modern-day arms race, artificial intelligence is driving how the countries wage wars in the future and the U.S. tool will not need human hackers in cyber warfare

Conventional warfare is slowly making way for a more sophisticated strategic attack and defense system. With the advent of artificial intelligence (AI), countries are changing the way their military works. China, Russia and the U.S. already have an AI-powered weapons system that may soon be deployed to defend and attack. But with more and more infrastructure getting linked to networks, cyber warfare is another tactic in the modern era that countries are adopting.

The cyber warfare units of North Korea, China and Iran have frequently been cited by the U.S. and its allies in recent times but Washington too has secretly developed its cyber unit that is powered by AI and would not require human hackers.

Age of Cyber Warfare

While unmanned drone attacks have become more mainstream for militaries around the world, the U.S. had launched a cyber attack on Iran. In 2005, in a joint operation with Israel, the U.S. targeted Iran's highly secured nuclear program with a malware named Stuxnet that halted the country's uranium production, albeit, temporarily. The cyber attack blew up centrifuges disrupting uranium enrichment.

Cyber Crime
Cyber warfare will be driven by AI rather than human hackers (representational image)

However, that was 15 years ago and there was no central command center. But that successful malware attack changed the country's stance. In 2009, the U.S. established a formal planning unit at the National Security Agency (NSA) headquarters at Fort Meade.

Project IKE

Since then the technology has evolved tremendously, forcing the U.S. to adapt to the changes like its adversaries. One aspect of that is automating the system of strikes and there is nothing better than AI. Hence, the Pentagon began testing machine learning that could not only defend its military assets linked to the network but could also launch attacks.

Thus, the Defense Advanced Research Projects Agency (DARPA) leaned toward automating the system with a project named Foundational Cyberwarfare which eventually became "Plan X" in 2012. Later, it became Project IKE with full implementation in 2018. The budget has also constantly increased to upgrade the system. Compared to this year's $27 million, next year, Pentagon will spend $30.6 million.

Artificial Intelligence
AI-powered Project IKE automates cyberwarfare (representational image) Pixabay

Data Analytics and Automation

It continuously assesses a target and predicts the success of a cyber mission with programs that adjust depending on the situation and information available. IKE being the brain, decides when and how to strike a target, effectively changing the way to wage a war, digitally. With no minimal human intervention, the system is fast, efficient and effective. Another advantage is that AI-powered hacks are difficult to predict as it consciously changes to adapt to the target.

The machine learning algorithm scans through heaps of data generated by the federal agencies and U.S. forces over the years. And the big data analytics helps the artificial brain to decide whether a strike is needed. It then churns out odds in percentage for commanders to decide whether to go on without requiring any human intervention.

"That was what was powerful. It categorized risk in a way that I could have a pretty good level of confidence," Ed Cardon, former head of the cyber forces (2013-2016) told Yahoo News. He further added that during Barack Obama's presidency, the decision to wage a cyber war, especially against ISIS, was difficult to take due to the fears of political repercussions. But with the system offering confidence, that is not the case anymore.

NSA headquarters
Project IKE is under US Cyber Command at the Fort Meade in Maryland where NSA also has its headquarter Wikimedia Commons

In 2013, former President Barack Obama issued a classified order that outlined steps needed to launch an attack after a high-level White House meeting, which was later amended removing the necessity to take approval from the Oval Office. Now, the Secretary of Defense could give the nod if the odds were high.

Potential Disadvantages

However, even such a sophisticated system has its shortcomings. For example, the source code of Stuxnet malware that was so successful eventually ended up on the internet, revealing how the U.S. launched the attack. In 2010, security researchers found the bug that exploited zero-day vulnerabilities in the Windows system.

With IKE, if its source code somehow ends up on the internet, it will cease to exist. Hence, any information about project IKE remains classified. But there are other aspects too. Computers can have glitches and if it takes a decision during those moments of malfunction, it could potentially kill even American troops if it gives the system a slight advantage to strike a target. Hence, there must be checks and bounds for automation in this modern-day arms race tool.

"The dangerous thing is that those numbers aren't always right. It's tempting to assume that, just because something came from a computer, it's rigorous and accurate," said Ben Buchanan, Cybersecurity and Foreign Policy Professor at Georgetown University.

Related topics : Artificial intelligence