It doesn't say what to watch for. And I think that's where the gap needs to be closed. It doesn't say, for instance, what the controls are monitoring for. Nowhere does the coordinate what would count as a model crossing a biosecurity line, who would decide if it did or what would happen next? And the key point here is an auditor must audit against something. And so we need to work to establish what those thresholds are. And there's some. (11:56–12:05)
STATEMENT
Asserts that the AI accord lacks clarity on biosecurity thresholds and decision-making processes.
This is an excerpt from a raw transcript. The assertions have not been audited or verified. Tap the video to view the source footage and understand the context.