Well it is already painfully visible that "checks & balances plus peer review" is barely managing to hold science together. Research has already been thoroughly gamed on a systemic level. Scientific funding and the whole "publish and perish" culture pretty much promotes bullshit over real research nowadays. And while solutions are easy to name (increase rigor for statistical significance, publish negative results, make scientist pre-publish description of planned data analysis before collecting data, etc.) they are almost impossible to implement, as everyone inside the system is trapped by the incentive structures that drive the whole thing.