Won't the scientist use some relatively secure/private model to fraud-check their own work before submitting? If it catches something, they would just improve t

Won't the scientist use some relatively secure/private model to fraud-check their own work before submitting? If it catches something, they would just improve the fraud.