After days of arduous negotiations, countries have agreed to begin working towards an international agreement on regulating lethal autonomous weapons systems.

AFP reported that talks at the United Nations' Review Conference of the Convention on Certain Conventional Weapons in Geneva, Switzerland, lasted five days and went into overtime before yielding a result in the early hours of Saturday (21 November).

Conference president Julien Thoni told reporters that the 128 states parties had agreed to begin drafting an international instrument on lethal autonomous weapons systems, with the work to be completed within the next two years.

Artificial intelligence (AI) is fundamentally transforming warfare. Non-governmental organisations had hoped for swifter, legally binding rules on the issue of handing decisions over human life and death to autonomous machines.

Although countries reached this consensus, the process remains fraught, with divisions between countries and a requirement for full consensus on any decision.

Thoni said the work ahead would clarify “which weapons can be forbidden according to international humanitarian law, and which is not forbidden but can be regulated”.

Campaign groups, however, said the outcome fell far short of expectations.

The Stop Killer Robots coalition said countries had failed to set out a clear path towards a legally binding instrument.

Verity Coyle, deputy arms director at Human Rights Watch, also expressed disappointment that, despite rapid advances in the technology, the conference had failed to spur the urgent action needed to establish new rules.