This lecture from the Instituto de Física Interdisciplinar y Sistemas Complejos (IFISC) explores how machine learning copies can contribute to developing more trustworthy artificial intelligence systems. Discover the theoretical frameworks and practical applications of creating ML copies as a strategy to enhance AI reliability, transparency, and safety. The 70-minute presentation delves into how these copies can serve as verification mechanisms, helping to validate AI behavior and identify potential issues before deployment in critical systems.