AWS updated its Service Terms on June 10, 2026 to add AWS FinOps Agent (Preview) to the list of AI Services and services that incorporate generative AI features. More substantially, AWS restructured and expanded its abuse detection provisions: it now requires explicit consent from users of certain Anthropic models for AWS to transfer user content and metadata to Anthropic for abuse detection processing, and it clarified that Bedrock stores inputs and outputs for up to 30 days solely for abuse detection purposes on identified models. This shifts abuse detection from an implicit AWS-only process to one that may involve third-party model providers with user consent.
Users: If you use an Anthropic model on Bedrock, you must opt-in to allow AWS to send your content to Anthropic for abuse detection.
Developers: AWS will store your Bedrock inputs and outputs for up to 30 days on certain models specifically for detecting policy violations.
The updated terms now require users of certain Anthropic models to provide explicit consent for AWS to transfer their content and associated metadata to Anthropic for abuse detection and processing. Previously, abuse detection was described as an AWS-only internal function. The revised terms clarify that Bedrock stores service inputs and outputs for up to 30 days on identified models solely for abuse detection purposes. Users of Anthropic models can provide this consent through the opt-in mechanism described in the applicable service documentation, as stated in section 50.12.2.2 of the updated terms.
→ Review whether you are using Anthropic models on AWS Bedrock; if so, evaluate the consent mechanism in your service documentation.
→ If deploying to end users, implement opt-in consent collection for Anthropic abuse detection and communicate the Anthropic Data Processing Addendum.
→ Review Anthropic's DPA and compare it against your existing data processing agreements and privacy frameworks.
ConductAtlas has recorded 7 material changes to this document over 31 days of monitoring (since May 2026). An additional minor or cosmetic changes were excluded.
3 of AWS Bedrock's significant changes have been classified as negative for consumers.
Users must now explicitly consent for AWS to transfer content and metadata to Anthropic for abuse detection processing.
Inputs and outputs are retained for up to 30 days on identified models solely for abuse detection, separated from service improvement uses.
This change record describes what was added, removed, or modified in the document. Analysis reflects what the updated agreement states or permits. It does not constitute a legal determination about enforceability. Applicability may vary by jurisdiction. Methodology
AWS restructured its abuse detection provisions on June 10, 2026 to require explicit user consent for certain Anthropic models to receive customer content and metadata. The change moves abuse detection from implicit AWS retention to …
Regulatory exposure, obligation change, escalation trigger, board-ready language, and recommended action for legal and compliance teams.
Unlock the full institutional analysis — InsightConductAtlas provides verified policy intelligence sourced directly from platform documents. All analysis is intended to support, not replace, legal and compliance review. Record CA-C-002819.
AWS Bedrock removed language requiring explicit consent to transfer customer content and metadata to Anthropic for abuse detection, and removed …
AWS expanded the list of services classified as 'Indemnified Generative AI Services' in its AWS Service Terms, adding three additional …
AWS Bedrock updated its Service Terms to introduce a new AWS (new) program and restructured billing and project management terms. …
Get the research letter
Companies change their terms quietly. We read every version and catch what actually changed. One email a week on the changes that matter and what they mean.
Get alerted when this policy changes again, including what changed and why it matters.