Coherent Extrapolated Vision

The challenge of AI adopting human morality lies in the complexity of determining which set of values and beliefs should be programmed into the AI. Different cultures have diverse moral standards, such as attitudes towards women's rights, dietary preferences, and sacred practices. To address this issue, a proposed solution involves creating a one-use superintelligence to identify the most universally acceptable values for ensuring human well-being and happiness. This concept, known as coherent extrapolated vision, entails letting the machine determine what values a super-intelligent machine should uphold, based on common and shared human values. The vision includes the idea of having the super-intelligent machine design a new machine aligned with this vision. While this approach presents technological challenges, it offers a potential blueprint for future AI development.