AI security guide · MITRE ATLAS

Model and AI service access management

How should access to AI models and services like Amazon Bedrock be controlled? AI services are cloud APIs. Who can invoke models, manage agents, change guardrails or read invocation logs is an IAM question — and misconfigured or public AI endpoints are a growing exposure.

ATLAS v2026.10Last updated
1. Reconnaissance12. Resource Development23. AI Attack Adaptation34. Initial Access (this technique)45. AI Model Access (this technique)56. Execution67. Persistence78. Privilege Escalation (this technique)89. Defense Evasion910. Credential Access1011. Discovery1112. Lateral Movement (this technique)1213. Collection1314. Command and Control (this technique)1415. Exfiltration1516. Impact16ATLAS5 OF 16 TACTICS
4. Initial Access5. AI Model Access8. Privilege Escalation12. Lateral Movement14. Command and Control

Why it matters

Unrestricted model access enables data leakage through prompts, abuse of paid capacity and tampering with safety configuration.

How it works

  • Stolen credentials used to invoke paid models.
  • Publicly exposed inference endpoints queried directly.
  • Guardrail or logging configuration modified by over-privileged identities.

Curated agent example

Fictional example: shared development keys

  1. A team shares one access key with bedrock:* for experimentation.
  2. The key leaks; an outsider uses it to run large volumes of model invocations.
  3. Costs spike and invocation logging is disabled.
  4. Mitigation: per-user roles, scoped bedrock permissions, protected logging configuration and cost alerts.

Fictional and simplified for education.

On AWS: architecture notes

  • Scope bedrock:InvokeModel to specific model ARNs.
  • Protect PutModelInvocationLoggingConfiguration and guardrail changes with SCPs or narrow permissions.
  • Monitor cost anomalies for AI services.
AWS IAM review bounds what a manipulated agent can do. It does not prove that the agent resists prompt injection.

MITRE ATLAS techniques

  • AML.T0040 AI Model Inference API Access

    AI Model Access · Maturity: Realized

    Adversaries may gain access to a model via legitimate access to the inference API. Inference API access can be a source of information to the adversary (Discover AI Model Ontology, Discover AI Model Family), a means of staging the attack (Verify Attack, Craft Adversarial Data), or for introducing data to the target system for Impact (Evade AI Model, Erode AI Model Integrity).

  • AML.T0096 AI Service API

    Command and Control · Maturity: Realized

    Adversaries may communicate using the API of an AI service on the victim's system. The adversary's commands to the victim system, and often the results, are embedded in the normal traffic of the AI service. An AI service API command and control channel is covert because the adversary's commands blend in with normal communications, so an adversary may use this technique to avoid detection. Using existing infrastructure on the victim's system allows the adversary to live off the land, further reducing their footprint.

  • AML.T0132 Misconfigured or Publicly Exposed AI Services

    Initial Access · Maturity: Demonstrated

    AI agents and LLM platforms are deployed across diverse architectures, including standalone servers, SaaS platforms, and low-code builders. All of these often have misconfigured access controls, ranging from missing authentication in LLM runtimes to permissive public access settings, that significantly expand their attack surface.

  • AML.T0012 Valid Accounts

    Initial Access · Privilege Escalation · Lateral Movement · Maturity: Realized

    Adversaries may obtain and abuse credentials of existing accounts as a means of gaining Initial Access. Credentials may take the form of usernames and passwords of individual user accounts or API keys that provide access to various AI resources and services. Compromised credentials may provide access to additional AI artifacts and allow the adversary to perform Discover AI Artifacts. Compromised credentials may also grant an adversary increased privileges such as write access to AI artifacts used during development or production.

ATLAS mitigations

  • AML.M0019 Control Access to AI Models and Data in Production

    Require users to verify their identities before accessing a production model. Require authentication for API endpoints and monitor production model queries to ensure compliance with usage policies and to prevent model misuse.

  • AML.M0005 Control Access to AI Models and Data at Rest

    Establish access controls on internal model registries and limit internal access to production models. Limit access to training data only to approved users.

  • AML.M0004 Limit AI Service Query Volume and Rate

    Limit the number and rate of requests that users can submit to an AI service. Apply limits by user, API key, tenant, device, or other authenticated identity. Use short-term rate, burst, and concurrency limits together with longer-term usage quotas.

  • AML.M0024 AI Telemetry Logging and Monitoring

    Implement logging and monitoring across the AI lifecycle to detect suspicious activity and support investigation. Capture security-relevant access, changes, and operations across data pipelines, model development, deployment, and connected systems. Preserve sufficient context to trace activity across components and identify coordinated behavior across accounts and access pathways.

Related Enterprise ATT&CK techniques

Controls to prioritize

  • Least-privilege identity policies

    Grant only the actions and resources a workload or person needs. Treat wildcard actions on IAM, STS, KMS, Secrets Manager, S3 and logging services as high-risk and review them explicitly.

    Security best practices in IAM
  • Auditable agent and model activity

    Record which user and agent triggered each tool call, the parameters, the result and the content sources used, alongside model invocation logs and CloudTrail.

    Monitor model invocation using CloudWatch Logs and Amazon S3
  • Cost and quota anomaly monitoring

    Unexpected compute or AI-service spend is often the first visible sign of hijacked resources. Monitor cost anomalies and keep service quotas close to real need.

    Detecting unusual spend with AWS Cost Anomaly Detection

Cloud Control Explorer →

Common misconceptions

“AI services are outside the IAM program.”
They are AWS APIs governed by the same IAM policies and logging.

Checklist

  • Inventory who can invoke and administer AI services
  • Enable invocation logging
  • Protect logging and guardrail config
  • Cost anomaly alerts

Run the AI Agent Security Self-Assessment

Related attack scenarios

More AI security guides

References

MITRE ATT&CK® and ATT&CK® are registered trademarks of The MITRE Corporation. ATT&CK data © 2026 The MITRE Corporation. This work is reproduced and distributed with the permission of The MITRE Corporation (ATT&CK v20.1). MITRE ATLAS™ data (v2026.10) © MITRE, used under the Apache License 2.0. Securitain is not affiliated with or endorsed by MITRE. Explanations, scenarios and control mappings on this page are Securitain’s own analysis. Sources and licensing

Stay in the loop

Cloud security insights, early access & compliance updates

Weekly InsightsEarly AccessExpert Content
No spam, everUnsubscribe anytimeWeekly updates
Chat with Securitain