Technically it's just lip service and nothing of the sort is ever going to happen any time soon, but it's still nice to see people speaking up
we can do two things at once
-
pass laws preventing companies from making LLMs that do things that LLMs categorically cannot do, but that LLM companies claim LLMs can do for marketing reasons
-
arrest CEOs who break that law (which will be difficult)
we do not need to, for example, consider any present or future societal impacts of AI (on the environment, on the labor market, on deskilling, on disinformation...) besides it one day getting Too Smart™ and deciding to kill us all, which we take it as given that it will be able to do