#WPSAdvice: Killer robots: a new frontier for Canada to advance WPS

With more than 120 active armed conflicts raging globally, opportunities to advance the Women, Peace and Security (WPS) agenda are shrinking as States and non-state actors alike forego comprehensive solutions to conflict, such as diplomacy and disarmament, in favour of military action and arms proliferation. Amid this widespread violence and heightened geopolitical tensions, emerging technologies and artificial intelligence (AI), including autonomy in weapons systems, are being rapidly developed and deployed in the use of force. As the new Canadian federal government aims to “[deploy] AI at scale” at home, it is crucial that emerging technologies are used to uphold a rights-based world order in both situations of armed conflict and in ‘peacetime’, rather than exacerbating inequality, injustice, and human suffering. 

What do killer robots have to do with WPS?

Autonomous weapons systems are the range of weapons systems that detect and apply force to a target based on sensor inputs, rather than an immediate human command. This means that after decision-making and activation by a person, there is a period when the weapon system can apply force to people or objects without any additional human approval: the specific object to be attacked, and the exact time and place of the attack, are determined by sensor processing, not by humans. While we shouldn’t panic about humanoid robots taking over the world as depicted in The Terminator or other sci-fi movies, we should be concerned by the real prospect that human control over the critical functions of selecting and attacking targets will be replaced by sensors, AI, and algorithms within weapons systems. 

Promoting and protecting the rights and agency of women, girls, and gender-diverse people experiencing armed conflict is at the heart of the WPS agenda. Weapons play a central role in the perpetuation of violence. But the connections between disarmament and arms control as measures to advance the WPS agenda remain underexplored in all three iterations of Canada’s National Action Plans on WPS. 

Tech is not neutral: risks of bias & discrimination

AI, and technology more broadly, are not neutral — the prejudices and inequalities present in our society also live in data sets, imagery, categories, labels, and algorithms. Canada has repeatedly raised concerns about the risks of algorithmic and other AI biases in the development and use phases of autonomous weapons systems, especially gender bias. Gender matters when it comes to the development, use, and governance of autonomous weapons systems – 92% of software professionals worldwide are men, women constitute between 0.3% to 20% of military personnel in most countries, while women represent only 1 in 3 diplomats in arms control and disarmament meetings. Policies and practices generated in male-dominated environments can overlook the complexities of intersectional identities, and in the context of autonomous weapons systems, there are significant risks that biases implicitly encoded in these systems could result in discriminatory outcomes, causing death or injury. In a 2021 review of 133 AI systems deployed across sectors, gender bias was exhibited in 44% of the systems, with over 1 in 4 of these systems also found to exhibit racial bias. In-built biases and other forms of discrimination present in our societies, which are then reproduced in target detection and force application of a weapons system, will inevitably result in disproportionate harm to those already marginalized by gender, race, disability, ethnicity, religion, or other identity qualifiers. 

Increasing autonomy in weapons systems risks the erosion of meaningful human control in the use of force and presents numerous human rights, humanitarian, technological, and ethical challenges. Empowering machines with the ability to autonomously kill people without meaningful human control is an egregious form of digital dehumanization, as complex human beings risk being reduced to mere data points. The consequences of algorithmic and other forms of AI bias are particularly acute in the context of autonomous weapons systems designed to target humans and underscore the urgent need for Canada to support the prohibition of such anti-personnel systems. 

The world has changed significantly since the Security Council passed resolution 1325 in 2000, and WPS priorities should adapt accordingly. We agree with Canada that “autonomous weapons systems may not be consistent with the principles related to the Women, Peace and Security agenda,”. The WPS agenda stands to support the critical role of women and girls in conflict prevention, resolution, and peacebuilding and autonomous weapons designed to target humans stand as a detrimental technology incompatible with this framework and its values. Viewing the governance of autonomous weapons through a WPS lens allows us to fully consider their impacts on diverse individuals in situations of armed conflict, as well as their impacts on human rights outside of conflict, given their potential to be used in policing and border enforcement operations. Integrating disarmament measures into WPS, particularly through the prohibition of anti-personnel autonomous weapons systems, would substantively reinforce all four pillars of the agenda by promoting human security and gender equality.

Our advice to the new government is to foreground disarmament in WPS initiatives and for the ‘responsible use of AI’ to begin with the negotiation of an international treaty on autonomous weapons systems that bans systems designed to target humans. Canada has a long history of championing human rights and disarmament, most famously demonstrated through the Ottawa Process that led to the Anti-Personnel Mine Ban Treaty. Canada has the opportunity to once again take a leadership role in disarmament by actively supporting the negotiation of a legally binding treaty prohibiting and regulating the use of ‘killer robots’ or autonomous weapons systems. New rules and norms on killer robots would not only reinforce existing legal frameworks such as international humanitarian law and international human rights law, but would additionally serve to strengthen the WPS agenda by ensuring that technology is used to promote principles of peace, justice, and human rights.

United Nations Secretary-General António Guterres states that “time is running out to take preventative action” on killer robots. Immense harm is already being perpetrated against civilians and other protected groups as a result of developments in AI and autonomy in the use of force, such as in Gaza and Ukraine, and legal red lines must be drawn to prevent further digital dehumanization. Canada should take Arnold Schwarzenegger’s famous “I’ll be back” line seriously when it comes to disarmament and lead once again by negotiating new international law regulating autonomous weapons and banning systems that target people before it is too late.

The views in this blog are those of the authors only and do not necessarily represent those of the WPSN-C or its membership.