Search Open Technical Databases
Description
Adversaries may search for publicly available research and technical documentation to learn how and where AI is used within a victim organization. The adversary can use this information to identify targets for attack, or to tailor an existing attack to make it more effective. Organizations often use open source model architectures trained on additional proprietary data in production. Knowledge of this underlying architecture allows the adversary to craft more realistic proxy models (Create Proxy AI Model). An adversary can search these resources for publications for authors employed at the victim organization.
Research and technical materials may exist as academic papers published in Journals and Conference Proceedings or stored in Pre-Print Repositories, as well as Technical Blogs. Publicly accessible AI services may be identified via Scan Databases.
- ATLAS ID
- AML.T0000
- Maturity
- RealizedUsed by real adversaries against AI systems in the wild.
- Platforms
- Enterprise
Sub-techniques
Mitigations
- AML.M0000Limit Public Release of Information
Limit the connection between publicly disclosed approaches and the data, models, and algorithms used in production.
- AML.M0039AI Honeypots
Decoy assets are catalogued by the same internet scan databases (e.g., Shodan, Censys) that adversaries query to locate exposed AI infrastructure. Adversarial engagement that originates from those platforms can grant defenders visibility into those discovery channels.
Case studies that use it
- AML.CS0001Botnet Domain Generation Algorithm (DGA) Detection Evasion
- AML.CS0003Bypassing Cylance's AI Malware Detection
- AML.CS0005Attack on Machine Translation Services
- AML.CS0007GPT-2 Model Replication
- AML.CS0010Microsoft Azure Service Disruption
- AML.CS0011Microsoft Edge AI Evasion
- AML.CS0012Face Identification System Evasion via Physical Countermeasures
- AML.CS0070Threat Actor Uses a DeepSeek-Powered Hermes Agent in Langflow and n8n Exploitation Attempts
Related ATT&CK technique
ATLAS adapts this technique from MITRE ATT&CK Enterprise.