“Hats off to Claude!”, the Don said. In an extraordinary, electric showing, The Machine was actually doing real (!), significant contribution in solving an open problem on Hamiltonian Cycle. If the Don has successfully utilized AI to solve an open problem, even skeptics start to wonder whether the future is truly inseparable from AI.

Whatever the future is like, though, one must admit that AI is destructive, however useful they are. In a lot of cases, AI amplifies the Dunning-Kruger effect for the user while confidently producing wrong responses that look deceptively smart. But, arguably, the tool is just misused by a lot of people¹. Another big concern is that the path towards getting the AI to the current state was not morally proper. It was riddled with wasteful energy consumption, questionable training sources², and bias.

While we have established that the training of AI agents is morally dubious, the question remains whether using AI is also morally wrong. I would like to argue that it very much depends on the future, and unfortunately, I am not an oracle.

Consider hypothetically that some researchers used AI to produce an end-all solution to global warming. Is the use of AI now justified? One can acknowledge the dark history behind AI, while also acknowledging the benefit of it. A parallel that can be taken in context of history would be the gruesome Unit 731 and the use of the experiment results. I repeat that we should not even think about repeating how we get those data. But now that the data is there that has a potential benefit to society, would you not use it?

1: See the use of AI by Don in the previous paragraph to show that AI can be correctly utilized.

2: Stolen.