OpenAI patents AI system to find security flaws in computer code
OpenAI has filed a patent for a system that uses artificial intelligence to find security weaknesses in computer software. The system first uses reverse-engineering tools to rank how risky a piece of code is, then uses a large language model to create a specific threat profile. Finally, the AI analyzes the code against that profile to identify potential vulnerabilities.
In the filing’s own words
The present disclosure relates to systems, non-transitory computer-readable media, and methods for improving the security analysis of computer code. In particular, the disclosed systems can retrieve a binary function from a binary repository of a computer application. In one or more embodiments, the disclosed systems can perform an initial threat analysis of the binary function by utilizing a reverse-engineering platform to determine a threat ranking for the binary function. In some cases, the disclosed systems can utilize a large language model to generate a threat model in the form of a structured data object that defines threat parameters corresponding to the computer application. Additionally, the disclosed systems can utilize a large language model and threat model to analyze the binary function based on the threat ranking to generate a threat object that indicates threat data associated with the binary function.
What exactly they claim (claim 1)
1. A computer-implemented method comprising: retrieving, by a generative artificial intelligence (AI) system, a binary function from a binary repository of a computer application; determining, utilizing a reverse-engineering platform to perform an initial threat analysis of the binary function, a threat ranking for the binary function; generating, utilizing a large language model of the generative AI system, a threat model comprising a structured data object defining threat parameters corresponding to the computer application; and generating, based on the threat ranking and by utilizing a large language model to analyze the binary function together with the threat model, a threat object indicating threat data associated with the binary function.
Facts
- Company
- OpenAI (OpenAI OpCo, LLC)
- Official title
- DETECTING VULNERABILITIES ASSOCIATED WITH BINARY FUNCTIONS UTILIZING GENERATIVE AGENTS
- Publication number
- US20260352690A1
- Application number
- 19365017
- Filed
- Oct 21, 2025
- Published
- Nov 5, 2026
- Inventors
- 3
- Claims
- 20
Classified under
H04L 63/1433 Electricity & electronicsMore from OpenAI in Security & privacy
- OpenAI patents AI agents that scan code changes for security flaws Nov 5, 2026
- OpenAI patents AI agents that find security flaws in code Nov 5, 2026
- OpenAI patents system that hides passwords from AI agents Nov 3, 2026
Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.
