Here comes the friendly, neighbourhood Terminator

Thomas Malthus famously worried about mankind, or sizeable chunks of it, at least, starving to death, as a result of population growth that exceeds humanity’s capacity to feed itself. Human ingenuity decimated the Malthusian prophecy, producing food far in excess of what humankind needs either directly or indirectly, that is, as animal feed that would be converted into meat, milk and eggs. However, we now confront the problem that human ingenuity produces machines with intelligence that humans can no longer control, and could, in theory, turn on humans.

Remotely controlled mine detectors have been in operation since the late 1960s, when the US forces used them in the Vietnam war. The 1970s saw the use of Explosive Ordnance Disposal (EOD) robots or, simply, bomb robots, in Northern Ireland. But those were primitive tools, surpassed in capability and sophistication by successive generations of advance in microelectronics, robotics and computing. US security forces have used robots extensively in Iraq and Afghanistan.

Under a government programme to sell army surplus to police forces, American police forces have acquired assorted ordnance across the country. That included robots. In 2015, a robot was sent up with a phone and a pizza on its arms to negotiate a would-be suicide off his mission. That was in San Jose. The next year, in Dallas, the police sent off an EOD robot with a mission to explode, rather than defuse, a bomb, to an assailant who was holed up with a gun that he had already used to shoot a dozen officers, five of them dead.

Robots don’t have feelings, at least, not yet. Set off a bomb, defuse a bomb, both are tasks, on which to report mission accomplished, if that is physically possible, that is.

Robots that can go where a human officer cannot without risking life and limb certainly offer a major advantage, a major power projection for the police. But, with great power comes great responsibility, even for those who don’t swing at one end of a strand of spider web. Policemen who can inflict lethal violence have to be held to account, in normal circumstances. When they can do so without the risk of any harm to themselves, that need becomes all the more imperative. Social identities and hierarchy could intrude into the calculus that goes into deploying lethal robots.

Robots in a human image are already a reality. With ever faster advances in artificial intelligence, it is not impossible to conceive of androids wielding weapons and relative autonomy, operating within given parameters. SkyNet and Terminator might still be some time away in the future, but robots that can kill are a reality. Robots that go rogue are a future nightmare, but rogues who use killer robots could be upon us.

There is no alternative to preparing for possible futures and for clear and present opportunities and threats.

India must encourage artificial intelligence research in all its forms: that is the only way to acquire the capability needed to retain strategic autonomy, as innovations raise artificial intelligence to the level of strategic capability. We have all seen nightmare videos of swarms of tiny flying drones that carry explosive charges blasting their way through walls, zeroing in on targets and blowing them up. Even when these drones are remotely controlled, the effect would be deadly, but when they are controlled by their own artificial intelligence, which can improvise tactics to meet a given objective, they would be deadlier still.

The reluctant voting in San Francisco Supervisory Council to permit deployment of lethal robots sounds the cold warning that those swarming drones are a reality in the making, not a nightmare from which we can wake up sweating.

We have to arm ourselves with technology and coherent ethical frameworks, within which to countenance robotic embodiments of new lethal capability, wielded on our behalf and against us.