There's no requirement to test if your ideas work if you're an academic. The funding pays for papers to get published, and that's all that's checked. To publish a paper you might need to at least pretend to test your ideas, maybe, but only in some of the better fields and journals. Nobody will check if you actually did test them though, so just making data up is a perfectly viable strategy. Worst case, you might get caught ten years from now if some random independent chooses to investigate and manages to go viral on social media. In very extreme cases you might get fired, but that won't stop you immediately getting another job doing exactly the same thing at another university (c.f. Brian Wansink).
But those are in the best, most scientific fields. In others you can just write papers all day that say all kinds of things, never test or even predict anything, and still have a successful career. That's how you get papers like the famous "feminist glaciology" paper [1], or thousands of COVID papers that present model outputs but never validate if the results matched reality, or thousands of string theory papers that don't make any predictions.
None of these problems will stop academics from being cited by journalists in high profile news outlets as if their ideas are already validated, nor being consulted by powerful politicians who then transcribe their policy demands into law, nor having their claims be automatically taken as gospel on forums like HN. If their ideas do become law and then cause a disaster, nothing will happen to them and they won't even suffer loss of reputation. For example, Einstein was a big supporter of Soviet-style planned economies, as were many academics of his era. That was a disastrous idea, and his supporting it - as late as 1949 no less - should have harmed his reputation for being a genius. But nobody remembers this today.
In these respects, they are the very epitome of what it means to be elite.
[1] https://journals.sagepub.com/doi/10.1177/0309132515623368