Bio-Evolutionary principles for general intelligence AI agents
Meshandren Naidoo · 2023
Evolution is an ecological phenomenon arising from the purposive engagement of organisms with their conditions of existence. This view presented here changes the paradigm of classical genetics evolutionary theory. It is not correct to separate evolutionary biology into processes of inheritance, development, selection, and mutation. Instead, the component processes of evolution are jointly caused by the organismal agency and also their ecological relations with their affordances. An understanding of the processes of evolution as being interlinked and widely dispersed shows that genes are unlikely to be the sole factor that controls evolution. There is no difference in kind between innate and acquired traits, or between novelties arising from genes and developmental or environmental sources. No criterion exists to differentiate between evolutionary and non-evolutionary events – an event is only evolutionary if it has an enduring effect on a lineage, which can only be determined retrospectively. The proper paradigm for generally intelligent artificially intelligent (AI) agent consideration, regulation and development requires a change in perspective at all levels. Firstly, these technologies should be modelled and treated as natural, purposive agents which enact their own evolution in pursuit of their goals. Secondly, given this shift in perspective, the development of technologies should focus on constraining mechanisms which ensure the stability of form through mechanisms like buffering through redundancy – or facilitated adaptive change. Furthermore, clarity on the issue of attribution should centre on questions of informational asymmetry.