g_pepper comments on OpenAI makes humanity less safe

[g_pepper]

Some AGI researchers use the notion of a utility function to define what an AI “wants” to happen. How does the notion of a utility function differ from the notion of a will?