Towards Blockchain-based AI Alignment Using a Proof of Personhood Consensus Mechanism

Recent developments in the field of Artificial Intelligence (AI) raised the possibility of timely emergence of a human-level artificial intelligence, and thus increased the urgency of AI safety. As AI alignment lags behind other developments in AI, this work explores how blockchain technology can support AI safety and how society can be involved in AI alignment. Other works propose the use of blockchain technology for the purpose of AI safety, but mostly without referring to novel consensus mechanisms, providing theoretical formulations or describing the potential societal impact of their implementation. This paper suggests to write AI alignment rules in a blockchain that can only be updated by humans which is ensured by the Proof of Personhood (PoP) consensus mechanism, provides a theoretical formulation of this approach and discusses its significance to society. Thus, relevant concepts are identified, combined and theoretically formulated to propose a system that is protected from AI interference and enables society to obtain and retain control over AI. The creation of such a system is of importance in an era where a human-level artificial intelligence could emerge that may one day reject human ethics, goals, and principles.