UN Solutions will fail because they only focus on LAWs, not other applications of AI – inflexibility limits potential solutions.
Kahn and Horowitz, 2021 – Research and Senior Fellows at the Council on Foreign Relations [Lauren and Michael, The Washington Quarterly 44:4 “Leading in Artificial Intelligence through Confidence Building Measures” https://doi.org/10.1080/0163660X.2021.2018794 Acc 6/6/22 TA]
One might argue that the United States should let others lead on AI, focusing instead on developing AI-enabled capabilities and not concerning itself with how other countries behave. But there is no substitute for American leadership and its ability to rally countries around the world to support shared standards. If promoting norms of responsible behavior with AI encourages other states to use military applications of AI in more responsible ways, it will create a more ethical and predictable security environment, likely benefiting the United States. Additionally, current international dialogue about military uses of AI focuses almost exclusively on lethal autonomous weapon systems (LAWS), the subject of a Group of Governmental Experts in the Convention on Certain Conventional Weapons.39 Currently, the international conversation has been largely been driven by NGOs such as the Campaign to Stop Killer Robots.40 While such conversations help bring attention to some of these issues, they oversimplify the risks and fixate on worst-case scenarios that are more likely outcomes of artificial general intelligence or human level machine intelligence rather than technology today. LAWS represent only a small fraction of the universe of potential issues surrounding military applications of AI. Broadening the international conversation about military uses of AI to incorporate the full scope of potential applications would generate better dialogue because it would include more of the real-world AI scenarios likely to confront militaries. Expanding the discussion would also allow states to pursue levels of control and regulation other than an all-or nothing ban and create a more calibrated and flexible range of approaches to different technologies with various levels and types of associated risks.