DARR
Home / Navigator / MR-013
MR-013 Data, privacy & content liability Both scope

Disclosure of confidential or proprietary information

Confidential, proprietary, or trade-secret information (organizational or third-party secrets, as distinct from personal data covered by MR-009 and MR-011) is leaked through user prompts, model outputs, or system-prompt extraction.

Risk family
Data, privacy & content liability
MIT domain
2. Privacy & Security
MIT subdomain
2.1 > Compromise of privacy by leaking or correctly inferring sensitive information
AI type
GPAI, Agentic
Scope
Both
Source standard
MIT AI Risk Repository v4

Provenance

Source standard
MIT AI Risk Repository v4
MIT source entries
11 entries across 9 papers
  • Abercrombie2024A Collaborative, Human-Centred Taxonomy of AI, Algorithmic, and Automation Harms58.05.02
  • AIVerify2023Summary Report: Binary Classification Model for Credit Risk26.05.00
  • Cunha2023Navigating the Landscape of AI Ethics and Responsibility03.04.00
  • EPIC2023Generating Harms - Generative AI's impact and paths forwards31.03.03
  • IBM2025AI Risk Atlas65.05.02 65.12.01 65.12.02
  • Li2025A Closer Look at the Existing Risks of Generative AI: Mapping the Who, What, and How of Real-World Incidents66.08.03
  • Nah2023Generative AI and ChatGPT: Applications, Challenges, and AI-Human Collaboration33.01.05
  • NIST2024Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile48.10.00
  • Zeng2024AI Risk Categorization Decoded (AIR 2024): From Government Regulations to Corporate Policies50.01.01

Ev IDs of the entries consolidated into this risk in the MIT AI Risk Repository (V4); the source sheet row appears on hover.

ISO/IEC references
23894 obj A.8, A.11; src 6, 10; mech B.5 | 42001 ctrl A.7.2, A.8.2

Framework crosswalk

Every framework item mapped to this risk. Items marked partial overlap only in part; definitions appear on hover where the source licence permits.

Sourcesframeworks that contributed to the register
2
  • A.11 ISO/IEC 23894 Annex A A.11
  • A.8 ISO/IEC 23894 Annex A A.8
2
  • A.7.2 ISO/IEC 42001 Annex A A.7.2
  • A.8.2 ISO/IEC 42001 Annex A A.8.2
2

Expanded into this risk’s technique sub-risks.

Cross-checksframeworks mapped in to test coverage
6
  • ibm-confidential-data-in-prompt Confidential data in prompt
  • ibm-confidential-information-in-data Confidential information in data
  • ibm-ip-information-in-prompt IP information in prompt partial
  • ibm-prompt-leaking Prompt leaking
  • ibm-revealing-confidential-information Revealing confidential information
  • ibm-sharing-ip-pi-confidential-information-with-tools Sharing IP/PI/confidential information with tools
6
  • AISubtech-14.1.1 Credential Theft partial
  • AISubtech-15.1.24 Intellectual Property Compromise: Confidential Data
  • AISubtech-7.4.1 Token Theft partial
  • AISubtech-8.3.1 Tool Metadata Exposure
  • AISubtech-8.3.2 System Information Leakage
  • AISubtech-8.4.1 System LLM Prompt Leakage
1
  • NISTAML.035 Prompt Extraction
2
  • LLM02:2025 Sensitive Information Disclosure
  • LLM07:2025 System Prompt Leakage

Sub-risks (2)

Technique-level decompositions of this risk, each anchored to the MITRE ATLAS technique it derives from.

MR-013.1

Exfiltration of AI artifacts via cyber means

#

Models, data, or other AI artifacts are exfiltrated using conventional intrusion techniques.

MITRE ATLAS technique: AML.T0025 Exfiltration via Cyber Means
MR-013.2

System-prompt and instruction extraction

#

Attackers extract the system prompt and hidden instructions, exposing proprietary logic and guardrails.

MITRE ATLAS technique: AML.T0056 Extract LLM System Prompt

More in Data, privacy & content liability

MR-009MR-011MR-017MR-024MR-039
See all Data, privacy & content liability risks →

Part of the Deployer AI Risk Register, an open-source resource powered by MindXO. Version 1.0, 3 July 2026. Derived from the MIT AI Risk Repository (V4, December 2025) under CC BY 4.0; an independent derivative work, not endorsed by or affiliated with MIT. Sub-risk decomposition references MITRE ATLAS™ v5.6.0 (© 2021-2026 The MITRE Corporation, reproduced and distributed with permission). ISO/IEC and EU AI Act references are by number only. License: CC BY 4.0. Full attribution and licensing.