There's nothing problematic about accepting the null hypothesis, it's just that instead of controlling for Type I error, you need to control for Type II error,

There's nothing problematic about accepting the null hypothesis, it's just that instead of controlling for Type I error, you need to control for Type II error, i.e. ensure sufficient power.