In their essay, the authors described a situation in which researchers had a debate about ethics … with AI. Artificial intelligence argued that “AI will never be ethical.” “This is a tool, and like any other tool, it is used both for good and for harm. There is no good AI, there are only good and bad people, ”his position is quoted.
The AI used was Nvidia’s Megatron Transformer, which was trained on a large amount of digital text. His answers are, of course, context sensitive. The team then asked the AI to express the opposite point of view as well. So the program “took” a position against AI.
As journalists note, the fact that Megatron Transformer can express opinions on both sides of the ethical debate simply confirms a more general idea: artificial intelligence can either follow the norms of human morality, or not, depending on the context.