Record summary
A quick snapshot of what this page covers.
Control summary
What this defense is meant to help prevent.
- ATLAS ID
- AML.M0004
- Priority score
- 80
Covered techniques
Attacks this defense is designed to help with.
AML.T0043.001 - Black-Box Optimization
Restricting the number of queries to the model limits or slows an adversary's ability to perform black-box optimization attacks.
AML.T0034 - Cost Harvesting
Limit the number of queries users can perform in a given interval to hinder an attacker's ability to send computationally expensive inputs
AML.T0043 - Craft Adversarial Data
Restricting the number of model queries can reduce an adversary's ability to refine and evaluate adversarial queries.
AML.T0005 - Create Proxy AI Model
Restricting the number of queries to the model decreases an adversary's ability to replicate an accurate proxy model.
AML.T0029 - Denial of AI Service
Limit the number of queries users can perform in a given interval to prevent a denial of service.
AML.T0014 - Discover AI Model Family
Limit the amount of information an attacker can learn about a model's ontology through API queries.
AML.T0013 - Discover AI Model Ontology
Limit the amount of information an attacker can learn about a model's ontology through API queries.
AML.T0062 - Discover LLM Hallucinations
Restricting number of model queries limits or slows an adversary's ability to identify possible hallucinations.
AML.T0024 - Exfiltration via AI Inference API
Limit the volume of API queries in a given period of time to regulate the amount and fidelity of potentially sensitive information an attacker can learn.
AML.T0024.002 - Extract AI Model
Limit the volume of API queries in a given period of time to regulate the amount and fidelity of potentially sensitive information an attacker can learn.
AML.T0024.000 - Infer Training Data Membership
Limit the volume of API queries in a given period of time to regulate the amount and fidelity of potentially sensitive information an attacker can learn.
AML.T0024.001 - Invert AI Model
Limit the volume of API queries in a given period of time to regulate the amount and fidelity of potentially sensitive information an attacker can learn.
AML.T0043.003 - Manual Modification
Restricting the number of model queries can reduce an adversary's ability to refine manually crafted adversarial inputs.
AML.T0046 - Spamming AI System with Chaff Data
Limit the number of queries users can perform in a given interval to protect the system from chaff data spam.
AML.T0005.001 - Train Proxy via Replication
Restricting the number of queries to the model decreases an adversary's ability to replicate an accurate proxy model.
AML.T0042 - Verify Attack
Restricting the number of queries to the model decreases an adversary's ability to verify the efficacy of an attack.
Source
Where this page information comes from.
Original source
Original source links
Open the public records and source datasets used for this page.