Ethical Ai? Yes Its Solvable and Trivially so

Ethical AI is a trivially solvable problem in (a) hardware (b) software design (c) requirement of insurance, and (d) extremely harsh punishment of violations of that law, applied to every person in the chain of decidability. (d) international treaty.

We have solved this problem for thousands of years among humans with one single rule. All civilizations and all law is based upon that one rule. That politicians, philosophers and theologians ‘skirt’ that rule does not mean we cannot apply it to software.

There is nothing ethical or moral about war. That war exists defines the limit of ethics and morality. There will be killing machines just as there are machine guns and nuclear weapons, and the first people to invent them will dominate war, politics, economics, for a century.

The military incentive always DEFINES the political order. Not the other way around. You cannot stop this technology. This tech means greatest manufacturing capacity and engineering capacity will dominate all future wars – and therefore politics and therefore economics.

However, it is entirely possible to protect citizens from criminal uses the same way we do from nuclear weapons. However, the cost of AI will be in the billions today and dependent on vast infrastructure. But this price will decrease while the cost of refining n-weapons won’t.