The purpose of prediction markets is to communicate insider knowledge.

Yes. And indeed, when aggregated and averaged across all betters, nobody makes any money.

The question isn't what percentage of bets resolve to no, but whether there is a consistent bias in the prices away from the fair price, which has an expected value of 0, and what direction that bias is in.

I hate that many people or even the news and scientists have already started to see the odds of prediction market as fact.

I'm sure in the near future, policy decisions or war strategies will be decided by prediction markets' odds, if they are not already being used.