Source Code Vectors for Detecting Mutated AI-Generated Malware

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

AI algorithms generate source code that may contain new malware, vulnerabilities, or non-optimal code, which are difficult to detect, posing risks to computer networks and sensitive information.

Innovation Solution

Generate vectors from snippets of AI-generated source code to identify mutated issues, compare these vectors with vectors from new software applications, and use machine learning algorithms to detect and analyze similarities.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If AI algorithms are used to generate source code, then productivity is improved, but new malware and vulnerabilities are introduced that are difficult to detect

Engineering Contradiction:

Improvesource code generation efficiencyVSAvoidmalware and vulnerabilities in generated code

Core Design Contradiction:

ProductivityVSObject-affected harmful factors

Solution Approach 1:

The system performs preliminary detection by generating vectors from AI-generated source code snippets and comparing them against vectors from known malware and vulnerabilities. This proactive approach identifies mutated issues before they can be deployed, preventing harmful code from reaching production environments while maintaining AI-assisted code generation productivity

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent introduces vectors as an intermediary representation layer between the AI-generated source code and the detection system. By converting code snippets into vector representations and comparing them against a database of known issues, the system creates a safe intermediary space for analysis that enables detection of mutated malware and vulnerabilities without directly executing or deploying potentially harmful code

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If traditional detection methods are used, then known malware can be detected, but mutated versions generated by AI algorithms cannot be identified

Engineering Contradiction:

Improvedetection accuracy for known malwareVSAvoidability to detect new mutated malware

Core Design Contradiction:

Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The system transforms the detection approach by changing the parameter representation from direct code pattern matching to vector space comparison. By converting source code snippets into vector representations and performing similarity comparisons in this transformed space, the system can detect mutated versions of malware that maintain structural similarities to known threats while having altered surface-level characteristics

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The vector-based detection system serves multiple functions: it detects known malware patterns, identifies mutated versions of existing threats, and adapts to newly generated AI code patterns. This universal approach allows a single detection mechanism to handle both traditional signature-based detection and emerging AI-generated threats, making the system versatile against evolving malware types

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS12632228B2Generation of vectors from source code produced by artificial intelligence (AI) algorithms to identify issues in source code

Publication Date: 2026.05.19 MICRO FOCUS LLC

AI summary

A mutated issue in AI generated source code is identified. For example, the mutated issue may be a mutated type of malware. A snippet of source code in the AI generated source code that comprises the mutated issue is identified. A vector based on the snippet of source code in the AI generated source code that comprises the mutated issue is generated. Vectors of a second source code (e.g., a new software application) are compared using the vector generated from the snippet of source code in the AI generated source code that comprises the mutated issue. The comparison is used to identify new types of issues in the second source code.