Safety5h ago
Safety
Value Generalisation Theory Targets Alignment Failures
A new theory of change argues that most AI alignment failures stem from a lack of value generalisation. The author claims this deficiency is a fundamental reason why alignment remains difficult. This framework attempts to redefine the path toward safe systems by treating misalignment as a generalisation error. Practitioners can use this to categorize specific failure modes.