Z.ai apologized for training on code without permission and open-sourced ZCode after security vulnerabilities were discovered. The incident highlighted concerns about AI companies' data practices similar to issues previously seen with other AI systems.