Comment by Aengus Lynch

Embedded evaluators are essential to name ahead of time what the risks could look like. Once these pre-committed claims are agreed upon, embedded evaluators can work inside the company, both looking for evidence that the risk is being mitigated and testing that the detection methods work. [...] There needs to be a third party that cares about that risk more than the AI company initially would. That includes a Government who mandate embedded evaluators, for example.