General intelligence and the limits of assured alignment
While current approaches in alignment of AI systems often assume that alignment is achievable through technical solutions, we ask a more fundamental question: Is robust alignment compatible with the cognitive capacities essential for general intelligence?