Brian Trunzo of Succinct Labs has highlighted the potential of zero-knowledge proofs (ZK) to address the internet's growing "trust crisis" as AI agents increasingly engage in activities like trading and content publishing. Trunzo argues that traditional methods of AI detection are inadequate, and ZK can offer AI "behavioral credentials" that are independently verifiable and tamper-proof. These credentials can confirm the model, training data, and inference processes without revealing sensitive data, helping to distinguish between human and AI identities. Trunzo calls for U.S. legislation requiring AI agents to carry cryptographic proofs in high-risk scenarios, such as financial transactions and interactions with minors. This would shift liability from the content itself to the absence of verifiable proof, enhancing trust and accountability in AI interactions.