Impossibility and Uncertainty Theorems in AI Value Alignment (or why your AGI should not have a utility function).

Peter Eckersley

Browse the full AAAI paper archive.