The International Committee of the Red Cross (ICRC) has called on governments to urgently push for a legally binding international treaty to control autonomous weapons, after the Seventh Review Conference of the Convention on Certain Conventional Weapons (CCW), held this week in Geneva (Switzerland), ended without setting a clear and immediate mandate to start negotiations.
The organization warned that the rapid development of artificial intelligence (AI) and its growing ability to act without direct supervision by developers, evaluators or users makes it increasingly necessary to set global limits for these systems. In this regard, it stressed that decisions about life and death should not be left in the hands of machines.
Although the ICRC considers that a significant opportunity has been missed, it underscored the broad support of most States for opening formal talks during the past week, something which, in its view, keeps open the possibility of articulating a response to the humanitarian, ethical and legal risks associated with these technologies.
For this reason, the organization has called for the demand of 163 countries included in a resolution approved last month by the First Committee of the United Nations General Assembly to be addressed.
According to the ICRC, the future instrument should ban autonomous systems that pose an unacceptable risk of violating International Humanitarian Law (IHL), especially those that are unpredictable and those that can select and directly attack people, while also setting strict limits for the rest.
Along the same lines, the ICRC president recalled last Friday that multilateral agreements have succeeded in prohibiting or restricting arms such as biological and chemical weapons, antipersonnel mines and cluster munitions, drawing a parallel with the current debate on autonomous weapons.
"It is not possible to allow autonomous weapons to be developed and proliferate without international standards and effective human control," he warned, while also warning that future generations will judge the response of current leaders to the risk that machines may kill people on their own. "The risks are too great. After years of debate, the time to act has come," he concluded.