Atrocity Alert No. 495: Artificial Intelligence and Climate Change

Atrocity Alert is a weekly publication by the Global Centre for the Responsibility to Protect highlighting situations where populations are at risk of, or are enduring, mass atrocity crimes.

AS ARTIFICIAL INTELLIGENCE ADVANCES, UN OFFICIALS WARN OF GROWING RISKS TO CIVILIANS

Recent warnings from artificial intelligence (AI) researchers and technology companies about the risks posed by increasingly autonomous systems have intensified calls for governments to strengthen global AI governance. While attention has focused on the potential future threat of AI systems acting beyond human control, AI is already being used by state and non-state actors in ways that can facilitate human rights violations, violations of International Humanitarian Law (IHL) and atrocity crimes.

In early September, a researcher at Anthropic resigned publicly over fears about the pace of AI development and potential for increasingly autonomous systems to cause devastating harm. The warning was followed by calls from AI companies and governments to “pace the frontier” of AI development and from UN officials for stronger safeguards. The UN Secretary-General and UN High Commissioner for Human Rights have emphasized the need for immediate global cooperation to confront AI-related risks, while the UN Security Council will hold a high-level open briefing on AI and international security today, 23 September.

The risks posed by AI are not solely about what these systems may do in the future. AI systems are already being utilized by state and non-state actors to exacerbate societal divisions and facilitate violations of human rights and IHL in ways that may amount to or increase the risk of atrocity crimes. AI is increasingly being used in warfare, particularly for target identification and decision-making, with potentially grave consequences for civilians. These uses raise significant questions regarding compliance with the IHL principles of distinction, proportionality and precaution.

On 6 July three people were killed in Zaporizhzhia, Ukraine, by a Russian drone equipped with an autonomous AI-guided targeting system. The drone reportedly autonomously selected its final target after being directed toward a gas station. Russia and Ukraine had already been using AI-enabled systems in drones before the emergence of increasingly autonomous systems.

Over the last three years Israeli forces have also reportedly used AI-assisted systems to conduct mass surveillance in Gaza and identify tens of thousands of people and critical infrastructure sites as potential targets prior to airstrikes. The United States (US) has similarly reportedly utilized AI to identify targets in its strikes on Iran. In September the UN Independent International Fact-Finding Mission on the Islamic Republic of Iran found “reasonable grounds to believe” that the US committed war crimes during two strikes in February that killed at least 178 civilians. This includes 123 schoolchildren killed by Tomahawk missile strikes on the clearly identifiable Shajareh Tayyebeh Primary School.

AI’s use in enhancing the precision of surveillance systems also raises serious concerns regarding privacy and digital security. This has increased governments’ capacity to target and repress certain populations, such as China’s use of AI-integrated cameras to surveil the Uyghur population.

On 15 September the UN Working Group on the use of mercenaries warned that AI is transforming the nature of mercenary and private military and security operations worldwide. The Working Group highlighted how these actors are increasingly relying on sophisticated technologies for surveillance and targeting, including against human rights defenders, journalists, ethnic and religious minorities and Indigenous Peoples. These actors can also access and control drones in conflict settings to conduct “technology-enabled attacks” and have reportedly been contracted to wage narrative warfare through “disinformation for hire” campaigns that utilize AI-manipulated content to influence perceptions of war and enable technology-facilitated gender-based violence.

As UN High Commissioner for Human Rights Volker Türk emphasized in his September statement on pacing AI development, “We need urgent action that is adapted to the unprecedented risks posed by frontier AI. At the same time, companies and States should not be allowed to sideline the need for strengthening existing regulations and accountability for AI-related harms from systems already in use.”

Governments must urgently work cooperatively to regulate AI and other emerging technologies through the lens of safeguarding human rights and protecting civilians. They must also hold AI developers accountable for their role in enabling violations and cease the utilization and transfer of technologies that facilitate human rights abuses and atrocity crimes. At the same time, confronting future threats requires AI developers to ensure that their systems and agents are designed with safeguards that protect human rights and ensure compliance with applicable IHL.

STRONG EL NIÑO COULD COMPOUND ATROCITY THREATS

Populations living in conflict-affected countries are among the most vulnerable to today’s climate crisis, with refugees and internally displaced populations disproportionately exposed to climate-related hazards. As global temperatures rise, extreme weather events, desertification and resource scarcity are intensifying the risks of hunger, poverty and conflict. Earlier this year, UN Secretary-General António Guterres stated, “A world in climate chaos cannot be a world at peace. Climate change is a threat multiplier…”

A rapidly intensifying El Niño is expected to significantly disrupt rainfall and temperature patterns across the world into early 2027. The World Meteorological Organization reports that El Niño is firmly established and expected to strengthen into a very strong event, increasing the risks of droughts, floods and extreme heat. “Disruption and devastation” from shifting rainfall patterns are already occurring and could worsen. The World Food Programme projects that climate-driven shocks associated with El Niño could push an additional 49 million people into acute food insecurity by the end of 2027, bringing the overall number to 274 million – an increase of 22 percent.

These impacts are already being felt in countries where populations face risk of, or are experiencing, atrocity crimes. In the Horn of Africa, El Niño-linked flooding threatens communities affected by conflict, displacement and severe food insecurity. According to the Danish Refugee Council, El Niño is expected to “supercharge” annual flooding during the rainy season in East Africa, potentially forcing up to three million people to flee by the end of this year. In the Central Sahel, El Niño is expected to increase agricultural drought and food insecurity, compounding existing conflict and atrocity risks. The Food and Agriculture Organization reports that food insecurity has deepened for five consecutive years in the Sahel.

In Myanmar (Burma), where 12.4 million people already face acute hunger and almost 4 million are displaced, greater rainfall variability, drought and flooding could further undermine food production and livelihoods among communities enduring war crimes and crimes against humanity. Meanwhile, Haiti faces heightened drought risks. During the 2015–2016 El Niño, harvests collapsed in Haiti by up to 70 percent in some areas, doubling food insecurity within months. Today, approximately 5.8 million people in Haiti face crisis levels of food insecurity driven by gang violence.

As world leaders gather in New York for Climate Week alongside the opening of the 81st session of the UN General Assembly, member states should treat climate resilience as an integral component of atrocity prevention. The international community must urgently support anticipatory action in countries facing overlapping climate, conflict and atrocity risks, including through food assistance, livelihood protection and locally led adaptation initiatives. Climate finance, adaptation strategies and early warning systems should integrate human rights, conflict sensitivity and atrocity risk analysis to identify and address risks before climate shocks intensify existing crises.