Someone's going to figure out that RFK's people used ChatGPT to write this report, I bet, and then no one checked it and it just made up fictional studies.

In my world policy decisions are built in facts. The references are part of the process by which you build confidence in the policy direction by showing the information on which the policy is based. If those references turn out to be faulty then the policy rationale falls apart.

But we are in the era of vibe governance and they don't care about facts. They will "fix" the document by selecting a new set of facts that matches the desired outcome. We used to refer to it as "decision based evidence making".