AI governance struggles with fragmented and compute-gated oversight

EU high-risk classification and GDPR debates, U.S. proposals on frontier AI and Pentagon incidents, OpenAI's proposed incident-disclosure response, and the “tokenocracy” critique point to unresolved accountability problems. Independent auditors may lack access, compute, or authority, while most reported legislative and regulatory measures remain proposals or early frameworks.