If you consider the combination of very fast improvements in AI, a lack of knowledge about abilities, high uncertainty about the future, the fact that guardrails are decided by AI labs, & that AI has very wide impact … expect mostly reactive, ad hoc & scattered policy responses.