Aligning Artificial Superintelligence via a Multi-Box Protocol
By: Avraham Yair Negozio
Potential Business Impact:
Makes super-smart AI agree on truth, not lies.
We propose a novel protocol for aligning artificial superintelligence (ASI) based on mutual verification among multiple isolated systems that self-modify to achieve alignment. The protocol operates by containing multiple diverse artificial superintelligences in strict isolation ("boxes"), with humans remaining entirely outside the system. Each superintelligence has no ability to communicate with humans and cannot communicate directly with other superintelligences. The only interaction possible is through an auditable submission interface accessible exclusively to the superintelligences themselves, through which they can: (1) submit alignment proofs with attested state snapshots, (2) validate or disprove other superintelligences' proofs, (3) request self-modifications, (4) approve or disapprove modification requests from others, (5) report hidden messages in submissions, and (6) confirm or refute hidden message reports. A reputation system incentivizes honest behavior, with reputation gained through correct evaluations and lost through incorrect ones. The key insight is that without direct communication channels, diverse superintelligences can only achieve consistent agreement by converging on objective truth rather than coordinating on deception. This naturally leads to what we call a "consistent group", essentially a truth-telling coalition that emerges because isolated systems cannot coordinate on lies but can independently recognize valid claims. Release from containment requires both high reputation and verification by multiple high-reputation superintelligences. While our approach requires substantial computational resources and does not address the creation of diverse artificial superintelligences, it provides a framework for leveraging peer verification among superintelligent systems to solve the alignment problem.
Similar Papers
Super Co-alignment of Human and AI for Sustainable Symbiotic Society
Artificial Intelligence
Makes super-smart AI learn good values with us.
Research on Superalignment Should Advance Now with Parallel Optimization of Competence and Conformity
Artificial Intelligence
Teaches super-smart AI to be good.
An International Agreement to Prevent the Premature Creation of Artificial Superintelligence
Computers and Society
Stops super-smart computers from being built too soon.