Need for meaningful human control over weapons highlighted by AI experts
For the governance of the use of force and the protection of people, every individual use of weapons should be subject to meaningful human control. This means a process of legal decision-making and deliberative moral reasoning by human beings for each individual attack. Weapons that do not allow such human control and attacks without such human control should be prohibited. As part of the Campaign to Stop Killer Robots, Article 36 argues that fully autonomous weapons should, as such, be pre-emptively banned.
In an open letter published by the Future of Life Institute highlighting recent advances in artificial intelligence (AI), more than 4,000 signatories including specialists from the field draw attention to the urgent need for research in order to ensure that AI systems “do what we want them to do” and maximise the benefits of this technology whilst avoiding “potential pitfalls”.
As well as detailing broader concerns around AI, the research priorities document attached to the letter gives attention to autonomous weapons and the need to articulate what ‘meaningful human control’ over weapons systems should look like, and how it can be maintained technically as systems develop. This theme has also been picked up in the considerable media coverage that the letter has generated (with its prominent signatories including physicist Stephen Hawking, tech entrepreneur Elon Musk and researchers from Google and Harvard, among others).
Within the framework of the Conventional on Certain Conventional Weapons (CCW) expert discussions on autonomous weapons systems, Article 36 has suggested ways in which debate and policy development could be structured around consideration of meaningful human control. We have called on governments to explain how control is understood and ensured for current weapons systems. This is essential not only to informing discussion on autonomous weapons and assertions about the lawfulness of future systems, but to evaluating current conduct.
Meaningful human control has already been a prominent feature of debate on autonomous weapons at the CCW and raised by United Nations Special Rapporteur on extrajudicial, summary or arbitrary executions Christof Heyns. The UN Institute for Disarmament Research has also analysed how a focus on meaningful human control may advance discussions. The attention being given to the concept at the CCW is encouraging, and we expect meaningful human control to be a subject of continued concern at the next expert meeting on autonomous weapons, to be held 13-17 April 2015. The acceptance of the concept of meaningful human control by states is vital to the discussion of the acceptability of different weapons systems.
Photo: Future of Life Institute
Posted in: Autonomous weapons,