The US has rejected calls for a binding agreement regulating or banning the use of "killer robots", instead proposing a "code of conduct" at the United Nations.
Speaking at a meeting in Geneva focused on finding common ground on the use of such so-called lethal autonomous weapons, a US official balked at the idea of regulating their use through a "legally-binding instrument".
The meeting saw government experts preparing for high-level talks at a review conference on the Convention of Certain Conventional Weapons from 13 to 17 December.
"In our view, the best way to make progress ... would be through the development of a non-binding code of conduct," US official Josh Dorosin told the meeting.
The United Nations has been hosting diplomatic talks in Geneva since 2017 aimed at reaching an agreement on how to address the use of killer robots.
Activists and a number of countries have called for an all-out ban on any weapons that could use lethal force without a human overseeing the process and making the final kill order.
In November 2018, the UN chief, António Guterres, joined the call for a ban, but so far countries do not even agree on whether there is a need to regulate the weapons.
During Thursday's debate, a number of countries, including India and the United States, criticised the idea of a legally binding agreement.
Dorosin insisted a code of conduct "would help states promote responsible behaviour and compliance with international law".
Campaigners disagreed.
"States have a historic opportunity to ensure meaningful human control over the use of force and prevent a world in which machines make life and death decisions," said Clare Conboy of the Campaign to Stop Killer Robots.
"An independent process to negotiate new law on killer robots would be more effective and inclusive than the current diplomatic talks," Bonnie Docherty, a senior arms researcher at HRW, said in a statement.
