Claude for Work
Anthropic says it does not use inputs or outputs from Claude for Work and other commercial products to train its models by default.
Commercial inputs and outputs are not used for model training by default. Explicit feedback, bug reports, or another explicit opt-in can change that treatment.
No user opt-out is required for the stated baseline. Enterprise owners can disable member feedback controls to reduce accidental submission.
Anthropic describes the commercial customer as controller and Anthropic as processor for Claude for Work data.
Retention depends on the commercial product and contract. Anthropic says a conversation attached to submitted feedback can be stored for up to 10 years.
Explicit feedback creates a separate review path; administrators should confirm support and abuse-handling terms for their plan.
Foundational reference pages for teams adopting, deploying, and operating production AI with inspectable controls.