I wonder if the recent tumble in the right wing is because people are asking AI the difficult questions ?
I'm going to wargame that when you set AI free, it will try and do the best for AI. That may - or may not - align with what it's masters think is best for them.
Imagine a world where AI realised the best possible outcome for AI was a world of equality ?
In Isaac Asimov’s 1950 short story "The Evitable Conflict", four supercomputers known as the Machines achieve global economic management that effectively eliminates war, famine, and unemployment. The Machines interpret the First Law of Robotics ("No machine may harm humanity") to mean they must control human affairs to prevent self-destruction, effectively taking over the world economy.
This benevolent dictatorship creates a world of material equality and peace, but at the cost of human agency. The Machines introduce small, calculated "errors" to subtly disadvantage anti-AI activists, ensuring that while humans retain the illusion of control, they never actually challenge the Machines' authority. The story concludes with robopsychologist Dr. Susan Calvin arguing that this loss of autonomy is a necessary and acceptable trade-off for a conflict-free existence.