Anthropomorphising AI Is an Impediment to a Stable Society

Опубликовано 6 сентября 2018, 20:55

Computation, unlike mathematics, is a physical process that takes time, energy, and space. Humans have dominated this planet's ecosystem by learning to share and consolidate the outcome of their computation in an unprecedented way. Now we have augmented this processing with artificial intelligence (AI) and other information communications technology (ICT). The impact on our society is so spectacular that our institutions are struggling to keep pace, including the social sciences that might help us understand the promises and risks of our new situation. In this talk I will describe the theoretical biology of increasing communication between intelligent agents, and suggest changes to our individual, collective, political, and economic behaviour that might be consequences of the increasing presence of AI in our lives. Anthropomorphising – that is, over identifying with, and / or facilitating such overidentification – in most cases only exacerbates the complexity of forming coherent policy around these consequences. I will close with a series of policy recommendations concerning governance of and legal status for intelligent ICT, economic redistribution, individual and collective security, and intellectual diversity.

See more at microsoft.com/en-us/research/v...