Versatile AI Risk Assessment

Source directoryBerryville Institute of Machine Learning (BIML)v1.0 (2020-01-13)

An Architectural Risk Analysis of Machine Learning Systems: Toward More Secure Machine Learning

“An Architectural Risk Analysis of Machine Learning Systems: Toward More Secure Machine Learning”. Publisher: Berryville Institute of Machine Learning (BIML). Version used: v1.0 (2020-01-13). The Versatile AI Risk Assessment threat catalogue backs 21 of its 52 threats with this document, at 24 verified locations.

As of: July 2026 · Catalogue version 2026.07.17.3

Document details

Publisher
Berryville Institute of Machine Learning (BIML)
Version used
v1.0 (2020-01-13)
Year
2020
Licence
CC BY-SA 3.0
Framework
BIML

Open the original document at the publisher External link; the publisher’s version always takes precedence.

References in the catalogue (24)

Every row states the clause, its title, the exact location in the document and the threats that refer to it.

Clauses of this document with the threats that reference them
Clause Title and location Referencing threats
BIML78 alg:1 Online PDF p. 16, [alg:1:online] Model Drift & Degradation
BIML78 alg:11 Parameters PDF p. 17, [alg:11:parameters] Supply Chain – Models
BIML78 assembly:2 Annotation PDF p. 13, [assembly:2:annotation] Targeted Poisoning / Label Poisoning
BIML78 assembly:6 Filter PDF p. 13, [assembly:6:filter] Model Reconnaissance
BIML78 data:1 Poisoning PDF p. 14, [data:1:poisoning] Training Data Poisoning
BIML78 data:2 Transfer PDF p. 14, [data:2:transfer] Supply Chain – Models
BIML78 data:6 Supervisor PDF p. 15, [data:6:supervisor] Targeted Poisoning / Label Poisoning
BIML78 eval:5 Catastrophic Forgetting PDF p. 18, [eval:5:catastrophic forgetting] Model Drift & Degradation
BIML78 inference:3 Confidence Scores PDF p. 20, [inference:3:confidence scores] Adversarial InputsModel Reconnaissance
BIML78 inference:4 Hosting PDF p. 20, [inference:4:hosting] Supply Chain – Infrastructure
BIML78 inference:5 User Risk PDF p. 20, [inference:5:user risk] Sensitive Information DisclosureShadow AI (Unsanctioned AI Service Use)
BIML78 input:1 Adversarial Examples PDF p. 19, [input:1:adversarial examples] Adversarial Inputs
BIML78 input:2 Controlled Input Stream PDF p. 19, [input:2:controlled input stream] Adversarial Inputs
BIML78 model:2 Trojan PDF p. 20, [model:2:Trojan] Backdoor ML Model
BIML78 model:5 Steal the Box PDF p. 20, [model:5:steal the box] Model Theft
BIML78 raw:1 Data Confidentiality PDF p. 10, [raw:1:data confidentiality] Privacy Attacks
BIML78 raw:2 Trustworthiness PDF p. 10, [raw:2:trustworthiness] Supply Chain – DatasetsTraining Data Poisoning
BIML78 raw:3 Storage PDF p. 10, [raw:3:storage] Supply Chain – InfrastructureExfiltration from ML Application
BIML78 system:1 Black Box Discrimination PDF p. 25, [system:1:black box discrimination] Hate Speech and Discrimination
BIML78 system:10 Denial of Service PDF p. 26, [system:10:denial of service] Model Denial of ServiceApplication Denial of Service
BIML78 system:2 Overconfidence PDF p. 25, [system:2:overconfidence] Overreliance
BIML78 system:5 Error Propagation PDF p. 25, [system:5:error propagation] Agentic AI / Autonomous Agents
BIML78 system:8 Insider PDF p. 26, [system:8:insider] Exfiltration from ML Application
BIML78 system:9 API Encoding PDF p. 26, [system:9:API encoding] Insecure Output HandlingApplication Vulnerabilities

This page does not reproduce the text of the standards. It states the identifier, title and location of the clause; the wording itself is in the original document. The mappings are taxonomic and not evidence of compliance.

Threats referencing this document (21)

Grouped by topic. Every entry leads to the full threat page.

Agentic and Autonomous AI

Application and Integration Security

Attacks on the Running Model and Service

Harmful Content

Model and Training Data Manipulation

Privacy and Data Leakage

Reliability and Responsible Use

Supply Chain and Provenance

More documents from the same publisher

All documents by Berryville Institute of Machine Learning (BIML) cited in the catalogue.

An Architectural Risk Analysis of Large Language Models: Applied Machine Learning Security v1.0 (2024-01-24) · 2024

From the reference to the assessment

The full catalogue states the mitigations, the possible impact and every verified location for each threat. The live demo runs locally in your browser, with no sign-up.

Go to the threat catalogue Start the live demo

Cite this page

For reports, policies or internal documents; the link leads directly to this source page.

“An Architectural Risk Analysis of Machine Learning Systems: Toward More Secure Machine Learning” (Berryville Institute of Machine Learning (BIML)). References in the AI threat catalogue, Versatile AI Risk Assessment, as of July 2026.
https://www.versatile-ai-risk-assessment.com/en/wissensbasis/sources/biml-architectural-risk-analysis-machine-learning/

← Back to the source directory