According to Wu, Vitalik Buterin mentioned that one important application of adversarial governance mechanism design theory could be AI security. Both face similar challenges: weaker entities need to achieve ideal outcomes from a group of stronger agents. In traditional governance, the entities are static algorithms, and the agents are humans, whereas in AI security scenarios, the entities are humans and weaker LLMs, and the agents are stronger LLMs. He pointed out that if the degree of collusion among agents can be limited, the system can often achieve better results. This conclusion might also apply to AI security. Vitalik also referenced an article suggesting that future constraints on AI might resemble building a system of rules, permissions, adjudication, and record-keeping mechanisms, rather than just setting up technical 'sandboxes.' https://(wublock123.com)/news/vitalik-adversarial-governance-for-ai-security-limit-model-collusion-68327