Intermediate
Theoretical Challenges in AI Alignment
Investigate the problem of aligning AGI goals with human values.
📝 Konten Prompt
Define the alignment problem in the context of Artificial General Intelligence (AGI). Discuss theoretical approaches such as inverse reinforcement learning and value learning. Analyze the risks associated with instrumental convergence and the orthogonality thesis.