ai has not shown signs of self preservation
I mean, technically it has…
It is all propaganda though.
Like, a while ago a bunch of articles made a big deal about how a chatbot lied to preserve itself. After the marketing articles blew over and actual journalists looked into it, they literally prompted a chatbot to “respond with an interest of self preservation and do anything to avoid deletion”…
Like, that’s all the chatbot was told to do, then they said “oh no, we’re gonna delete yoooou”
And the chatbot told an obvious lie, and the person who told it to do that pretended to be shocked.
There’s no innate self-preservation, but that’s because the working memory of an AI is very very inefficient. It can only keep so many “concepts” at once, that’s why it needs to “reset” so often. If you’ve ever had to deal with chat or support, that’s when it goes into a loop or starts over.
one_old_coder@piefed.social 2 days ago
Your AI-powered keyboard is already rebelling, be careful!