Free beta access. 50 free AI tokens each week. Optional top-ups available. See what is included
Opening Word document…

A lack of human control undermines accountability, which entangles NATO in conflicts initiated by AI.

Hill and Marsan, 2018 - Director and Senior Assistant, NATO Office of Legal Affairs [Steven and Nadia, 7-18-18 “Artificial Intelligence and Accountability: A Multinational Legal Perspective” https://www.sto.nato.int/publications/STO%20Meeting%20Proceedings/STO-MP-IST-160/MP-IST-160-PP-4.pdf Acc 4/21/22 TA]

4.0 ACCOUNTABILITY OF AI-ENABLED TECHNOLOGY International and domestic law offers some tools to cope with the effects of AI-enabled weapons from a multinational collective defense perspective. New technology does not necessarily require new laws and it is not the aim of this paper to propose the creation of new legal frameworks. The applicable legal framework will depend on the technology used and on the actual and potential effects caused by that technology and can include various national laws, human rights law, the law of state responsibility, international humanitarian law and the law of armed conflict. With machines taking on the qualities of human intelligence, including perception, cognition and action, the issue of accountability for illegal acts performed by autonomous systems has dominated the debate. The issue is not so much what legal framework applies, but who is legally responsible for the effects of AI-enabled weapons. Accountability and Responsibility AI-enabled action that constitutes an “armed attack” within the sense of Article 51 of the UN Charter, could potentially set off the invocation of Article 5 of the Washington Treaty. Setting aside the issue of the threshold to be applied to determine whether an AI-enabled attack would constitute an “armed attack”, the preliminary question that would arise is that of attribution. After all, assigning responsibility for an armed attack is a necessary precondition to the use of self-defence measures. AI-enabled weapons challenge our traditional notions of responsibility: who can be held accountable for the effects of this technology, especially when fully autonomous? Can we even talk of a “legal personality” of AI-enabled systems? In this area, States have been rather clear in expressing the need for responsibility to be attributed to human beings.18 The international debates and discussions on the legal approach to take with respect to lethal autonomous weapons systems (LAWS) provides a useful reference on the concept of “human control” where the substance of the debate has been how to regulate human-machine interaction, often referred to as “human-machine teaming”. The human role in independent machine decision-making can vary, as exemplified through the OODA loop of decision-making (Observe, Orient, Decide and Act).19 An “in the loop” system requires human intervention for its operation, an “on the loop” system provides for human intervention if needed, and an “out of the loop” system does not require human intervention at all, a prospect that is criticized by observers who have supported the concept of “minimum human control.” Privileging human judgment and accountability above mechanical efficiency, the United States, for example, has taken a clear stance towards asserting that a human being should always be kept in the decision-making loop for the use of LAWS, as the ultimate decider on the use of lethal force in the battlefield.

Sign in to recut, update, or save this card.