Old version
May 30, 2026 01:19 UTC
d60e9a71487307f8b041f5d0a466d87958df3404718291d4cb2b49f6b2cdec81
CA-V-003185
New version
June 2, 2026 01:09 UTC
fe2ad8e475c7223be304443c27fca9f43ce70ae337fb8c79c8f36229294fb926
CA-V-003289
Share 𝕏 Share in Share
Change Summary
AWS Bedrock updated its abuse detection and monitoring practices on June 2, 2026. The revised terms clarify that abuse detection mechanisms may now store service inputs and outputs for up to 30 days to detect policy violations, and specify that if child sexual abuse material is detected, AWS may review flagged content before reporting it to authorities. The prior language described automated detection without explicitly stating the 30-day retention period or the review step.
medium severity
0 Sentences added
0 Sentences removed
6 Sentences modified
1007 Sentences before
1007 Sentences after
Added
Removed
Modified
BeforeAfter
0AWS Service Terms Last Updated: May 29, 2026 1.0AWS Service Terms Last Updated: June 1, 2026 1.
45Certain Services may incorporate generative AI features, powered by Amazon Bedrock, that enable you to use prompts to generate output, including: Amazon Bio Discovery, Amazon CloudWatch, Amazon CodeCatalyst, Amazon Connect Customer, Amazon Connect Decisions, Amazon Connect Health, Amazon Connect Talent, AWS Database Migration Service, Amazon DataZone, Amazon Lex, Amazon OpenSearch Service, Amazon Personalize, Amazon Q, AWS Entity Resolution, AWS Transform, AWS AppFabric, AWS HealthScribe, AWS Resilience Hub, AWS App Studio, Kiro, Amazon SageMaker Data Agent, the model customization agent in SageMaker AI, Amazon Quick, AWS DevOps Agent, AWS Security Agent, Amazon WorkSpaces, and AWS Deadline Cloud. 1.24.1. The Amazon Bedrock automated abuse detection mechanisms may apply to such services.45Certain Services may incorporate generative AI features, powered by Amazon Bedrock, that enable you to use prompts to generate output, including: Amazon Bio Discovery, Amazon CloudWatch, Amazon CodeCatalyst, Amazon Connect Customer, Amazon Connect Decisions, Amazon Connect Health, Amazon Connect Talent, AWS Database Migration Service, Amazon DataZone, Amazon Lex, Amazon OpenSearch Service, Amazon Personalize, Amazon Q, AWS Entity Resolution, AWS Transform, AWS AppFabric, AWS HealthScribe, AWS Resilience Hub, AWS App Studio, Kiro, Amazon SageMaker Data Agent, the model customization agent in SageMaker AI, Amazon Quick, AWS DevOps Agent, AWS Security Agent, Amazon WorkSpaces, and AWS Deadline Cloud. 1.24.1. Amazon Bedrock abuse detection mechanisms may apply to such Services.
46See here for more details. 1.24.2. To improve performance, such services may use cross-region inference, using the optimal AWS Region to process your Content when running model inference.46See here and applicable Service-specific documentation for more details. 1.24.2. To improve performance, such services may use cross-region inference, using the optimal AWS Region to process your Content when running model inference.
556Notwithstanding anything to the contrary in the Agreement or Service Terms, Section 20 (AWS Marketplace) above applies to your use of models sold by third parties, and Amazon Web Services, Inc. is the invoicing party with respect to these models. 50.12.2. As part of providing the Service, Amazon Bedrock may use automated abuse detection mechanisms designed to detect harmful content, including related to potential violations of our or third-party model providers’ terms of service or acceptable use policies.556Notwithstanding anything to the contrary in the Agreement or Service Terms, Section 20 (AWS Marketplace) above applies to your use of models sold by third parties, and Amazon Web Services, Inc. is the invoicing party with respect to these models. 50.12.2. As part of providing the Service, Amazon Bedrock may store Service inputs and outputs for up to 30 days (unless otherwise required by law) for the sole purpose of detecting activity that violates our, or third-party model providers’, terms of service or use policies.
557If these mechanisms detect apparent child sexual abuse material, you agree and instruct that we may report the incident to the National Center for Missing and Exploited Children or other authority.557If we detect apparent child sexual abuse material (“CSAM”), you agree and instruct that we may: (a) review the flagged input or output to determine if it is CSAM, and (b) report the incident to the National Center for Missing and Exploited Children or other authority.
558See here for more details about abuse detection on Bedrock and related information sharing with third parties. 50.12.3. Provisioned throughput commitments for Bedrock.558See here for more details about abuse detection on Bedrock. 50.12.3. Provisioned throughput commitments for Bedrock.
Stay ahead of the changes

Watch this before it changes again

Follow unlimited companies, monitor the clauses that matter across every platform, and get the full institutional analysis on what each change obligates you to do.