Formulate a comprehensive alignment strategy for an Artificial General Intelligence (AGI) operating in an open-ended environment. Your strategy must address the 'paperclip maximizer' problem and incorporate inverse reinforcement learning techniques. Explain how you would mathematically formalize human values to prevent unintended consequences during the optimization process.