Indeed so, but the scientific process (in the general, Enlightenment sense) is more robust and abuse / fault -tolerant than anything else we've come up with.
Reproducibility crises or statistical hacking news articles are evidence of success. In a less-introspective system, those self-reevaluations didn't even happen!
Viz, the 300+ years it took for a shift from a geocentric to a heliocentric consensus.