Train multiple models for AI classification for Google Drive on demand

Google is enhancing AI classification in Google Drive to improve the efficiency of training custom AI models. A significant update is the introduction of multi-model support, allowing administrators to train up to five unique models for various labeling needs. Previously, only a single model per label field was supported. The training process is now on-demand, giving administrators control over when models are trained, replacing the former predefined schedule. This change allows organizations to manage their training pace interactively. Additionally, a refreshed user interface has been implemented, featuring a new onboarding flow and a model details page. This redesigned UI offers administrators richer insights into model training status, data metrics, recall scores, and version history. These updates are primarily targeted at Google Workspace administrators. The improvements aim to empower organizations to standardize data classification and achieve labeling consistency at scale. The AI classification capabilities enable automatic file identification, classification, and labeling within Google Drive. These labels can then be integrated with data loss prevention, lifecycle management, and auditing policies for enhanced data governance.