More not long ago, in 2022, OpenAI released its approach to the alignment challenge, anticipating that aligning AGI to human values would most likely be more difficult than aligning latest AI devices: "Unaligned AGI https://chatgptxopenai.com/login