Non-approved AI provider model in active use
- Severity
- Medium
- Service
- Bedrock
- Check ID
- BEDROCK_SHADOW_AI
What this check finds
A Bedrock batch inference job or prompt router references a foundation model from a provider not on the approved list (Amazon, Anthropic, AI21, Cohere, Meta, Stability AI, Mistral AI). Unapproved providers may not meet your organization's data-handling, licensing, or compliance requirements.
Passing looks like: Only approved AI models in use.
How to fix it
Review the flagged model provider. If intentional, add the provider to your approved list. Otherwise, update batch jobs and prompt routers to use approved providers only. Bedrock console → Batch inference / Prompt routers.
Compliance controls it is evidence for
A failing result counts against these controls in KloudLytics; a passing one is evidence towards them. How compliance mapping works
| Framework | Controls |
|---|---|
| PCI-DSS v4.0.1 |
|
| HIPAA Security Rule |
|
| SOC 2 — Trust Services Criteria |
|
| NIST SP 800-53 Rev5 (Moderate) |
|
| NIST Cybersecurity Framework 2.0 |
|
| ISO/IEC 27001:2022 Annex A |
|
AI risk mappings
| Framework | Controls |
|---|---|
| OWASP Top 10 for LLM Applications (2025) |
|
| MITRE ATLAS (Adversarial Threat Landscape for AI Systems) |
|
Checked on every scan
KloudLytics runs this check each time it scans a connected AWS account, through a read-only role, and lists every affected resource with its region. On Pro and Business a fix is written for the specific resource rather than the general case above. The exact access it needs
More Bedrock checks
- CriticalBedrock agent IAM role has broad permissions
- CriticalBedrock API key belongs to an administrator
- HighBedrock agent has no guardrail applied
- HighBedrock knowledge base not encrypted with CMK
- HighBedrock model invocation logging not enabled
- HighExternal model weights imported into Bedrock without governance review
- HighLong-lived Bedrock API key in use
- HighNo AWS Budget alert configured for Bedrock spend