A starting point for discussion
The idea
The tentative design gives two bodies that check the AI executive the power to pause it if it causes harm. It gives them separate administrators and separate credentials. Their practical independence may also depend on whether they share other tasks or rely on the same systems or people.
What this means and why it matters
What this means
Separate administrators are one part of independence. The bodies that check the AI executive may also need separate ways to get information and carry out decisions.
Why it matters
The proposal does not establish that every function must be separate. It asks which shared functions could give one actor leverage over both bodies that check the AI executive.
Conditions & assumptions
Which tasks and shared arrangements are safe remains unresolved.
Claims behind this perspective
Open a claim to examine its reasoning, evidence and objections. A claim can support more than one perspective.
No claims have been explicitly connected to this perspective yet.
Objections & tensions
AI-suggested relationships. Authors have not necessarily responded to one another.
Sources & further reading
No sources have been attached to this idea. Its reasoning should be examined on that basis.