Since most research articles have poor statistical power and researchers have freedom to choose among analysis methods to get favorable results, while most tested hypotheses are false and most true hypotheses correspond to very small effects, we are mathematically determined to get a plethora of false positives.
