Samar Ansari

The hardware-level mechanisms most crucial for verifiable AI treaty enforcement, like on-chip compute metering and cryptographic proof-of-training, are also the least technically mature, creating a critical gap in AI governance capabilities.