Nothing is ever “proven”. There is simply more or less support for a theory or proposition.

Replication and meta-analysis are an important part of this.

Most scientists are in fact very conservative with how they claim their results - less so university PR departments and “study shows” clickbaiters.