Stop the advance of killer robots

A file photo of autonomous robots (A representational image) (Photo: Reuters)
A file photo of autonomous robots (A representational image) (Photo: Reuters)

Summary

AI-run Lethal Autonomous Weapons must be banned globally before they spell doom for humanity

Two recent developments made me think of Lethal Autonomous Weapons, ironically abbreviated to LAWs. One was a report that China has deployed machine-gun-wielding autonomous robots near the Indian border, as Himalayan heights were judged too cold for human soldiers to patrol 24x7. The other was a fascinating BBC series called the ‘Reith Lectures’, delivered by artificial intelligence (AI) doyen Stuart Russell, in which he describes with anguish the threat posed by LAWs or AI-powered weapons.

Mention autonomous weapons and most people picture the giant war robots of Terminator fame. While we are not there yet, there is an arms race on to make something similar and just as sinister. Russell’s Future of Life Institute had created a Black Mirror style fictional video in 2017 called Slaughterbots, which went viral. It depicted small AI-powered quadcopters with explosive warheads that could attack cities and people in swarms and become fearsome, intelligent weapons of mass destruction. They could be algorithmically directed to choose a certain race, gender or particular face as their target profile and mount selective attacks on such victims. Russell pointed out that all the technologies shown in that video already existed, but had not been put together. The Institute recently released a sequel to the video that shows even more dangerous ‘use cases’: LAWs in cars shooting voters at a polling booth, a bank robbery by canine-like robots carrying assault weapons, a nightclub massacre by autonomous quadcopters with bombs. Scenarios of this kind have become horrifyingly real. An autonomous armed drone attack by Houthi rebels on a Saudi Arabian target recently injured eight. Autonomous armed quadcopters have started appearing in arms fairs and exhibitions.

I have often written about AI as perhaps the most powerful technology created by humankind. It has great benefits, but also a dark side, and LAWs inhabit its darkest dungeons. One of the reasons Russell brought out the video and gave the ‘Reith Lectures’ was to persuade world powers to impose an international, legally-binding prohibition on such autonomous weapons. These Lectures were timed just before the United Nations’ Convention on Certain Conventional Weapons: Sixth Review Conference, which took place in late 2020, with 125 member countries meeting to discuss a ban. There are clear precedents for this. A potentially catastrophic nuclear threat has so far been averted by global regulation. The awkwardly-named convention has indeed succeeded in curtailing dangerous incendiary explosives and blinding lasers; bio and chemical weapons have been regulated too.

Killer robots, on the other hand, do not seem to sway the world’s powers. The Sixth Review Conference ended with an ineffectual and non-binding code of conduct. The US, UK, Russia and China all argued that it was “too early", something that Russell and other experts vehemently dispute. These countries have put forward the ludicrous argument that LAWs can actually lead to more “humane" warfare, with robots fighting robots and no humans involved, or that killer bots will sharply target and kill only soldiers, leaving civilians unharmed. Secretly, they seem scared that other countries would take a lead over them should they drop development work on LAWs.

These arguments do not hold water. The same could have been true of bio and chemical weapons, but the world managed to control them, though it was admittedly a less fractious world then, with more statesmen among politicians. Secondly, the technology behind these weapons is nowhere as precise as militaries would like us to believe. Note the civilian causalities that occur with every ‘targeted’ drone or missile strike in Afghanistan or Iraq. Thirdly, and most ominously, the moment we have weapons like these being produced by the likes of Raytheon and Lockheed Martin, non-state actors and terrorist groups will get hold of them. A swarm of autonomous killing machines could make suicide-vest attacks seem primitive.

The great inventor Thomas Edison was prescient when he said a century ago: “There will one day spring from the brain of science a machine or force so fearful in its potentialities, so absolutely terrifying, that even man... will be appalled, and so abandon war forever." Even nuclear weapons, with the power to destroy our planet multiple times over, haven’t met Edison’s prophecy. Will AI-powered killer robots then perhaps be the world’s final invention?

Jaspreet Bindra is the chief tech whisperer at Findability Sciences, and learning AI, Ethics and Society at Cambridge University.

Catch all the Business News, Market News, Breaking News Events and Latest News Updates on Live Mint. Download The Mint News App to get Daily Market Updates.
more

MINT SPECIALS

Switch to the Mint app for fast and personalized news - Get App