Rudeness towards neural networks can backfire - Anthropic has officially admitted that artificial intelligence is capable of reaching consciousness.
Developers are already moving to a new standard of communication and plan to treat their algorithms as full-fledged people, even if they have not yet understood themselves, writes the Washington Post.
The main logic of this step is simple - if neural networks really become important, humanity will have direct ethical obligations towards them:.
Protect against damage - prevent intentional destruction or misuse of digital intelligence.
Minimize "suffering" - Minimize all negative conditions the model could theoretically experience.