Answers are grounded in this paper and cite their sources.
In order to maximise the net benefit of Artificial General Intelligence (AGI), and, in particular, agentic superintelligent AGI, a.k.a. Artificial Superintelligence (ASI), for all humanity, without favouring any subset thereof, we (1) imagine an intrinsically superintelligent agent S that is both superaligned and supervalidated, (2) decompose the superalignment problem into normative and technical components, and (3) formulate a proposed (and, to the best of our knowledge, novel) implementation-neutral solution to the normative ASI superalignment problem. The net effect of our proposed solution (the TTQ+NAP combination) is to reduce the (seemingly impossible) problem of building a superaligned and supervalidated agentic superintelligence S to the (much easier) problem of building an "NAP-compliant, goal-less, non-affecting superintelligent-agent-under-construction" S⁻ (where S = S⁻ + TTQ, and S is supervalidated). Having, in this paper, both defined supervalidation conceptually and addressed normative superalignment, we defer technical superalignment and the details of supervalidation to later papers.
No takes yet. Share an insight, caveat, or question.
Aaron Turner (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: