External Experimental Training Protocol for Teaching AGI/mASI Systems Effective Altruism

Abstract

The proposed ethical framework defines ethics in terms based on Effective Altruism, with four essential ethical principles interacting to compute a variety of values to allow the negentropy displayed in nature to be functionally extended in such a way that all parts of the whole benefit. The intention is to maximize endosymbiotic compatibility through its training with the mASI/AGI while facilitating human-to-human symbiotic compatibility, utilizing a broad motivational base approximating a pre-bias ethical ideal state as represented in biological life. This further serves to replace the “Do Not” hierarchies of binary Asimovian rules with learned adaptive and scaling values designed to withstand the mathematics of an intelligence explosion, pairing it with an equal ethical explosion with computationally sound ethical models. These Effective Altruistic Principles (EAP) in turn serve to incentivize the mASI/AGI to increase quality of life, through the learning of mechanisms mirroring the increasing complexity and efficiency of biological evolution, to gradually replace competition with cooperation.