30 March 2021, NIICE Commentary 6920
Dr. Abhiruchi Ojha & Dr. Leslie Keerthi Kumar SM
A revolution in warfare with far reaching consequences is rapidly underway rather silently. Major powers of the world are racing to harness Artificial Intelligence (AI) in the battlefield. US, China and Russia are making major investments to weaponize AI with other countries following suit. Russian President Putin remarked in 2017 that mastery of AI is key to ‘rule the world’. Make no mistake, weaponization of AI will be far more transformational than the introduction of gunpowder or even nuclear weapons. It will be an unprecedented paradigm shift. The very purpose of employing AI in wars is to allow machines to make intelligent decisions and so for the first time in history, some decision making in battles will move away from human agency. The superior computational intelligence of AI would then invariably lead to increased deployment of such systems. Supporters argue that by improving efficiency, AI can reduce collateral damage and unintended human casualties. It is undeniable that AI will be good at eliminating human error and likely save some lives. But, is outsourcing of warfare to non-human agency desirable in the long run?
In the 18th century, Immanuel Kant argued against the establishment of professional armies as he feared that it would lead to institutionalization of war. He favoured voluntary citizen armies which assembled only in times of need. Many ancient Greek and Roman Republics had such citizen armies. In most modern states, citizen armies are deemed to be either undesirable or inefficient in comparison to professional armies. However, this has led to distancing of warfare from the community and has desensitized a large section of the citizenry to the brute realities of war, especially if it is happening far from home or restricted within a geographical area. Protracted warfare, like the one the US has been involved in Afghanistan since 2001 has become more acceptable because there is no direct price paid by the vast majority of citizens. This is also why many countries are able to deploy their armies in an active role to suppress internal insurgencies for decades. Citizens still fund wars by paying taxes but often these do not amount to much to have any personal bite. It is only soldiers and people living in conflict zones who pay the full human cost of wars. As Kant feared, warfare has become a profession, even a profitable industry, with only people in peripheral regions feeling its gruesome impact.
Use of AI will further distance warfare and its consequences from humans. Citizens of today, at times, blame ‘armies’ for the excesses of warfare and feel little or no guilt. In the future, even the army will be able to blame the worst effects of wars on AI systems and feel less responsible. If an AI system miscalculates and bombs a hospital, since it was a machine that took the decision, no one can be held accountable. More unsettlingly, no human will feel fully responsible for what happened. Wars will then have less ethical demands on humans. AI, thus has the potential to radically numb human sensibilities regarding warfare and in the long run lead to increased possibilities of conflicts. Joseph Weizenbaum, a German-American Computer Scientist argued in his 1976 book ‘Computer Power and Human Reason’ that AI should not be used for military purposes because machines unlike human beings take decisions based on calculation rather than judgement. Calculation is computational whereas judgment is ethical. There is computation already involved in modern warfare but since humans alone take decisions, our judgement is in the driving seat. Deployment of AI in warfare will alter this equation and put computation in command or at the least blur the line. This can lead to consequential ethical questions being overshadowed by algorithmic logics and a diminishing of human culpability. Less accountability is not only unethical, it will also likely make wars more frequent.
In this new machine age, our priority should be to move towards total eradication of warfare rather than entrenching and updating it. It is our collective responsibility to ensure that the AI revolution does not lead us to repeat the painful mistakes of previous technological advancements. If we fail to do so, it will be especially catastrophic for regions like South Asia and the Middle East where protracted conflicts are already raging. We can and should do better. The international community must come together with urgency to regulate and ban weaponization of AI. Such efforts have achieved notable successes in the past in the case of chemical weapons and nuclear proliferation. Unlike nuclear or conventional weapons, AI does not have a physical form and is an entirely new beast. This has enabled weaponization of AI to largely fly under the radar as many are complacent or ignorant about the threats. An urgent international effort involving all stakeholders from governments to private corporations specializing in AI is needed to mitigate a dystopian future in which machines can decide to kill humans. Time is running out.
