It sounds like you're claiming that literally all predictive models just take some initial sampling bias and amplify it over time. Am I reading this right?
Especially if you use data created as a direct product of that predictive model to further tune that model.
That bias would accumulate.