Define the concept of the Alignment Problem in artificial intelligence. Discuss the theoretical difficulties involved in specifying a utility function that captures the complexity of human values without leading to unintended consequences (instrumental convergence). Provide examples of potential misalignment.