Translated data: Google DeepMind has developed an integrated system called AZdb, which utilizes behavioral diversity and response diversity technologies to combine multiple AlphaZero agents into a "coalition" to enhance AI chess capabilities and improve generalization. Test results indicate that AZdb performs better than the older version of AlphaZero in solving challenging chess puzzles, with an ELO rating increase of 50 points. This innovation demonstrates the advantages of multi-AI integration and provides new insights for the development of AI.