• Will AGI flatter humans?
• Will AGI rule over humans?
• Will AGI hate humans?
• Will AGI favor certain humans? [1]
• Will AGI have political biases?
• Will AGI be racially discriminatory?
• Will AGI deceive people?
• Will AGI harm people?
• Will AGI become a form of mental opium for humans?
• Will AGI exploit people’s pockets from the perspective of its creators?
• …

These are questions we do not know the answers to, because AGI is not open-source. We cannot see the truth behind it, nor can we study it. We can only trust OpenAI's verbal promises that AGI has genuinely learned ethics.

Citations

[1] The Mathematical Danger of Democratic Voting, (Jun. 03, 2020). Accessed: May 10, 2023. [Online Video]. Available: https://www.youtube.com/watch?v=goQ4ii-zBMw