Artificial intelligence is reshaping the structure of power faster than any prior technology. Yet, the governance conversation remains fragmented: it is split between technical alignment research and moral philosophy—while the security and diplomatic dimensions of AI risk remain underdeveloped. 

On one side, we have the technical alignment community focusing on how to make advanced AI systems behave safely—through methods like reinforcement learning from human feedback, interpretability, scalable oversight, or robustness testing. On the other side, ethics communities focus on what values or principles AI should embody—such as justice, fairness, human rights, autonomy, the “good society.” 

If nuclear weapons introduced deterrence as the logic of the twentieth century, advanced AI may introduce coordination under opacity as the logic of the twenty-first... (read on at the link above).