The article discusses many of my fave topics, but not as clearly as I'd like. It should be more clear in distinguishing ethics and morality which is used interchangeably by layfolk but shouldn't here.
The other thing it doesn't connect is how we currently use AI. What we use are statistical models not axiomatic proofs. We can extract axiomatic proofs from them, but that's an extra unexplained step which is needed to make the rest of the article relevant.
The conclusion is also the same as before we had AI, we'll still have to make moral judgements where ethical evaluations are inconclusive or unreliable. Pretty much the same as any safe use of AI not specific to this area.
Ironically it could have been been beneficial to use an AI to proof the article for strength of its argument and logical gaps.
The other thing it doesn't connect is how we currently use AI. What we use are statistical models not axiomatic proofs. We can extract axiomatic proofs from them, but that's an extra unexplained step which is needed to make the rest of the article relevant.
The conclusion is also the same as before we had AI, we'll still have to make moral judgements where ethical evaluations are inconclusive or unreliable. Pretty much the same as any safe use of AI not specific to this area.
Ironically it could have been been beneficial to use an AI to proof the article for strength of its argument and logical gaps.