TechCrunch AI's coverage of Hassabis's proposal highlights a pivotal tension in AI governance: who gets to set the rules. While the call for a technically-focused, independent body addresses the glaring deficiencies of recent ad-hoc government reviews, it arguably creates a new set of conflicts. A FINRA-like model funded by the very labs it regulates risks becoming a captive institution, potentially prioritizing commercial speed and industry consensus over public safety.
The plan's reliance on voluntary initial cooperation and its vague path to "formalisation" may leave it without teeth precisely when it needs them most. The real test will be whether such a body could ever say 'no' to a powerful lab's flagship model, a scenario the current framework seems designed to avoid.