Thanks to the rapid advances made in the field of AI, autonomous weapons systems, or killer robots in colloquial terms, might soon become a reality. As a result of the fact that this is the case, the UN has adopted a resolution to make these systems less effective than might have been the case otherwise. These types of weapons can acquire targets without any human involvement whatsoever, which makes them an especially dangerous outcome of the current AI race.
With all of that having been said and now out of the way, it is important to note that Harvard law lecturer Bonnie Docherty recently spoke out about this issue. She described autonomous weapon systems as systems that rely on sensor inputs to determine targets rather than human input. It turns out that they have actually been used multiple times in the past, although they are not quite as sophisticated as they would eventually end up becoming.
Systems used during the ethnic conflict in the Nagorno-Karabakh region were able to identify targets all on their own. The same can be said of the systems deployed in the Libya conflict, with some referring to them as loitering munitions. These weapons can hover over the field of battle and deploy their payloads as soon as an enemy target is detected, even if a human didn’t order the strike.
Needless to say, autonomous weapon systems come with a whole host of ethical concerns with all things having been considered and taken into account. It can reduce the taking of human life to a matter of numbers and data, which many consider to be crossing a line.
Algorithmic bias is also essential to consider because of the fact that this is the sort of thing that could potentially end up discriminating against people based on their ethnicity, gender and other aspects. Even disabled individuals could end up being targeted, with the AI based targeting systems unable to discern human rights in the appropriate circumstances.
Apart from ethical considerations, legal concerns have also arisen. Machines might not be able to differentiate between military combatants and humans that are present on the battlefield in a civilian capacity. Human judgement is essential in this regard due to the reason that weighing civilian casualties against military outcomes.
This involves something called the proportionality test, wherein someone or the other determines whether or not civilian loss of life justifies military action. For all of its advancement, AI can’t yet be programmed to display human judgement.
This raises another important question that must be asked. If the AI can’t show judgement, how can it be held accountable for any potential atrocities or crimes against humanity? At the same time, the operator of the system can’t be held accountable either, since they’re not technically the one that ordered the attack.
So far, any attempts to ban autonomous weapon systems have met stiff resistance from countries like Russia. Even the US and the UK have proposed non-binding resolutions in order to leave the door open for future use of these systems should the need arise. Indeed, a number of countries prefer non-binding resolutions, with each of them coincidentally developing autonomous weapon systems of their own.
As it currently stands, the UN is trying to collect civil opinions on the matter at hand. 164 member states voted in favor of this resolution, and it will be interesting to see where things go from here on out. According to the UN Secretary General, a new treaty might be coming as early as 2026. If it fails to reach the required number of voters, the potential loss of life might be staggering. Unlike landmines and other munitions, these aren’t tried and tested weapons yet, which might make obtaining a vote harder in the long run.
Photo: Digital Information World - AIgen
Read next: AI Incidents Increased by 30% Year Over Year
With all of that having been said and now out of the way, it is important to note that Harvard law lecturer Bonnie Docherty recently spoke out about this issue. She described autonomous weapon systems as systems that rely on sensor inputs to determine targets rather than human input. It turns out that they have actually been used multiple times in the past, although they are not quite as sophisticated as they would eventually end up becoming.
Systems used during the ethnic conflict in the Nagorno-Karabakh region were able to identify targets all on their own. The same can be said of the systems deployed in the Libya conflict, with some referring to them as loitering munitions. These weapons can hover over the field of battle and deploy their payloads as soon as an enemy target is detected, even if a human didn’t order the strike.
Needless to say, autonomous weapon systems come with a whole host of ethical concerns with all things having been considered and taken into account. It can reduce the taking of human life to a matter of numbers and data, which many consider to be crossing a line.
Algorithmic bias is also essential to consider because of the fact that this is the sort of thing that could potentially end up discriminating against people based on their ethnicity, gender and other aspects. Even disabled individuals could end up being targeted, with the AI based targeting systems unable to discern human rights in the appropriate circumstances.
Apart from ethical considerations, legal concerns have also arisen. Machines might not be able to differentiate between military combatants and humans that are present on the battlefield in a civilian capacity. Human judgement is essential in this regard due to the reason that weighing civilian casualties against military outcomes.
This involves something called the proportionality test, wherein someone or the other determines whether or not civilian loss of life justifies military action. For all of its advancement, AI can’t yet be programmed to display human judgement.
This raises another important question that must be asked. If the AI can’t show judgement, how can it be held accountable for any potential atrocities or crimes against humanity? At the same time, the operator of the system can’t be held accountable either, since they’re not technically the one that ordered the attack.
So far, any attempts to ban autonomous weapon systems have met stiff resistance from countries like Russia. Even the US and the UK have proposed non-binding resolutions in order to leave the door open for future use of these systems should the need arise. Indeed, a number of countries prefer non-binding resolutions, with each of them coincidentally developing autonomous weapon systems of their own.
As it currently stands, the UN is trying to collect civil opinions on the matter at hand. 164 member states voted in favor of this resolution, and it will be interesting to see where things go from here on out. According to the UN Secretary General, a new treaty might be coming as early as 2026. If it fails to reach the required number of voters, the potential loss of life might be staggering. Unlike landmines and other munitions, these aren’t tried and tested weapons yet, which might make obtaining a vote harder in the long run.
Read next: AI Incidents Increased by 30% Year Over Year
