The policy debate on artificial intelligence (AI) in the military domain increasingly emphasizes integrating international humanitarian law (IHL) into military AI system design, aiming for "lawful by design" capabilities. However, Laura Bruun and Netta Goussac from SIPRI argue that this focus risks states over-relying on industry for IHL compliance.
States must avoid unrealistic expectations of industry, which often lacks IHL expertise, as evidenced by its absence on company websites like Anduril or Palantir. Furthermore, IHL compliance cannot be solely engineered; it depends on the system's use and circumstances throughout its lifecycle, not just design. Crucially, IHL obligations primarily rest with states, not industry. Therefore, states must maintain independent capacity to assess and verify supplier claims through technical literacy, independent testing, legal reviews, or contractual requirements, ensuring "lawful by design" strengthens IHL compliance rather than diluting it.
No comments:
Post a Comment