Thesis of this case:
AI is inevitable. It is a normal part of modernizing our military. Because it is a dual use technology, it will diffuse much more easily, which means there is not a huge military advantage to be the First to develop it. So, we are not in an “Arms Race.”
However, we – along with China and Russia - believe that we are in an arms race. We believe we have to develop it first because if China develops it first, we will be at a huge disadvantage. Not true, but that is what we believe. China believes the same thing about us. This is what is called a “Security Dilemma” – our pursuit of our security undermines another nation’s security, and their reaction is to pursue their security in a way that undermines ours, without a natural stopping point. We “create” a race by believing in it, which is a self-fulfilling prophecy.
So, what is the harm of believing that we are in an arms race? This Arms Race Framing or Arms Race Mentality causes us to deploy AI weapons as fast as we can, which causes us to rush to deploy them before they are tested, evaluated, verified and validated (TEVV). Before they are safe and reliable. If we rush, we cut corners, and put out “brittle” AI, which means AI that collapses in the real world when the real world is different from their testing environment. It’s like when EA pushed out Battlefront II way too early to try to beat other game publishers to the market, and it was full of bugs and malfunctions. Except this time, “bugs” in military AI that controls our nuclear weapons could malfunction in ways that cause nuclear war.
So, the plan enhances testing, evaluation, verification and validation (TEVV) processes with NATO for all our AI weapons. This is also called Assurance. It adopts “life cycle” testing, which just means that you continuously test the AI over its time in use, because it evolves over time, due to machine learning. This will make sure that the weapons we put out work as designed, and don’t break down at inopportune times. There is some pretty good evidence that a large commitment to AI TEVV can build confidence globally, reducing the pressure to “race” to deploy weapons.
The advantages are just the different types of AI systems that could break down due to bugs, malfunctions or premature deployment. Nuclear Systems, Cyber Systems, Brittle Systems, etc. There is an Operator Trust scenario, which just says that soldiers won’t Trust their AI weapons if they are not safe and reliable, and that if soldiers and commanders do not Trust their weapons or systems, that breaks down the entire military readiness of NATO – blah blah blah – Russia/Ukraine/escalation something something nuclear war. You can choose one, two, or three of these scenarios to run. Probably not four, but hey – you do you.
This case is well positioned against security kritiks, because it addresses IR from a largely constructivist viewpoint explicitly in the 1AC. It is also well positioned to kritik the impacts to disads that rely on “You lose the Arms Race” arguments as being threat construction.