A recent interview between Quanta Magazine and Professor Judea Pearl, a pioneer of AI who has had significant contributions to AI development shed light on how Pearl presumes casuals reasoning could provide machines with human-level intelligence. Machines in the future could even “achieve status as moral entities with a capacity for free will—and for evil,” he believes.
In this interview, Prof. Pearl emphasized the importance in AI development of studying cause and effect, which he argues science has long neglected. He also wrote about similar concerns in his latest book titled The Book of Why: The New Science of Cause and Effect. He explained that since there is lack of the asymmetric language required to capture understanding causality, it prevents people from fully taking advantage of AI’s potential. Pearl believes that, in the future, machines could possess human-like intuition about cause and effect. This would raise the possibility of robots having a kind of free will, along with which comes the capacity for evil. For example, robots might “ignore the advice of other components that are maintaining norms of behavior that have been programmed into them”.
MDI has created the Seven Layers of AIWS framework, which guides AI development so as to promote positive outcomes and reduce risks associated with AI. Layer 1 involves designing a responsible code of conduct for AI Citizens to ensure that AI is safely integrated into human society. Layer 2 specifies an appropriate ethical behavior framework for AI Citizens. The Seven Layers were presented and sent to G7 Summit in 2018.