I mean the second one only works if you’re talking about climate change, which I’m for, but also we need to tall about corporate pollution as a whole.

Otherwise, AI systems are about as dangerous as what you give them permissions to.

Gotta ask how well maintained where the IT systems of the smaller organizations that where recently targets by “unaligned” corporate systems…

It wouldn’t shock me if an environment haphazardly thrown together by some vc skids who don’t care how anything is configured so long as it “just werkz” got pwned by an LLM.

The only other issue is hooking em up to weapon systems, however you could argue that’s just as stupid and dangerous as letting a predefined algorithm make the same decisions.