Not about 'killer robots': Britain must exploit competitive edge in AI, says Army cyber chief

Artificial Intelligence isn't about 'killer robots' and Britain would be 'mad' not to try and be a world leader in the technology, the UK’s military's cyber chief has said.

General Sir Patrick Sanders, the head of Britain’s Strategic Command, said artificial intelligence (AI) will be central to emerging technologies such as quantum computing, biotechnology and the military’s use of cyberspace. 

Speaking exclusively to the Telegraph in Estonia as he visited Nato forces, Gen Sanders said: “If you crack AI...you can master the mechanism to exploit and enhance technology at the pace you need.”

Noting that Britain is the third ranked cyber power in the world and a global leader in AI, he said “it would be madness if the UK didn’t try to exploit these”.

"Of all the new technologies, the one ring to rule them all is AI."

Machine learning, whereby computers learn through exposure to data to spot patterns and make decisions with minimal human intervention, is of particular interest to the armed forces.

“It gives you the opportunity for better decision making, it gives you the opportunity to begin to develop autonomy,” Gen Sanders said. 

“There’s a lot of concern out there about killer robots and ethics. Actually the real use of AI is to support humans, to be under command of humans. 

“The idea of human-machine teaming implies you can team with a computer. There isn’t a team, the humans are in charge, but AI can relieve the human of the burden of pouring through gigabytes of data to spot a disturbance on the road where there wasn’t one before [if a roadside bomb had been planted, for example]. 

“That ability to spot patterns and present that data to humans moves humans up the value chain.”

So-called ‘narrow AI’ is used every day in search engines or vehicle management systems that warn if a car is straying out of lane, for example. Unlike narrow AI that only really presents humans different courses of action, ‘general AI’ will decide its own objectives. 

Gen Sanders says routine life does not yet use ‘general AI’ and to consider doing so raises many questions.

“The ethics debate is important,” he said, adding it is important to guard against in-built human bias to such systems from the programmers that create them.

 “It is there to support humans and, ultimately, humans [must] have the whip hand.” 

Colonel Jaak Tarien, Director of the Nato Cooperative Cyber Defence Centre of Excellence, a Nato-accredited think tank, agrees, but warns: “Cyber defence is still full of human decisions”.

In developing AI systems to conduct cyber attacks, he says Russia and China “are not bothered by ethics, they want efficiency and power”.

“Ten percent of collateral damage is acceptable [to them]. For us zero percent is the desire. That makes it very expensive and complex for us," he told the Telegraph. 

“We will be struggling with our self-imposed rules but we have to do it. For the free world there is no other way. 

“Every single human being’s rights and privacy are important to us.”