I don't think we have a better, readily available solution to the problem than a public peer network (such as https://pubpeer.com/) monitoring the published papers, especially those "groundbreaking breakthroughs". Also, AI algorithms should be applied for the purpose. Every research funded from public funds should be freely available, together with datasets and auxiliary data (scripts, etc). Study preregistration should be obligatory.
Coordination of the scientific project is also very important, but that's out of my competence.
I don't think we have a better, readily available solution to the problem than a public peer network (such as https://pubpeer.com/) monitoring the published papers, especially those "groundbreaking breakthroughs". Also, AI algorithms should be applied for the purpose. Every research funded from public funds should be freely available, together with datasets and auxiliary data (scripts, etc). Study preregistration should be obligatory.
Coordination of the scientific project is also very important, but that's out of my competence.