NvidiaPatent applicationSecurity & privacyInterest 6/10

Nvidia patents system to score security risks in AI models

Patent drawing
Patent drawing from the official filing.

Nvidia describes a method for checking the security of artificial intelligence programs before they are used. The system scans the code to find specific features, runs tests in a safe environment, and calculates a score showing how likely the model is to cause damage if hacked.

Why it mattersThis is interesting because it suggests a way to measure the safety of AI systems, which is a growing concern as these models become more powerful and integrated into critical infrastructure.

In the filing’s own words

Devices, systems, and techniques for assessing and managing network security risks associated with an artificial intelligence (AI) model executed by a primary computing system. The techniques include receiving one or more executable files implementing the AI model. The techniques include scanning the one or more executable files implementing the AI model to identify a target attribute of the AI model. The techniques include executing one or more tests in a testing environment associated with a computing system hosting the AI model to generate test data associated with the target attribute. The techniques include generating, based on the test data, a damage likelihood score associated with the target attribute of the AI model.

What exactly they claim (claim 1)

1. A system comprising: a memory device; and at least one processing device, coupled to the memory device, to perform operations comprising: receiving one or more executable files implementing an artificial intelligence (AI) model; scanning the one or more executable files implementing the AI model to identify a target attribute of the AI model; executing one or more tests in a testing environment associated with a computing system hosting the AI model to generate test data associated with the target attribute; and generating, based on the test data, a damage likelihood score associated with the target attribute of the AI model.

Facts

Company
Nvidia (NVIDIA Corporation)
Official title
DETERMINING SECURITY RISK RELATED DAMAGE LIKELIHOOD ASSOCIATED WITH ARTIFICIAL INTELLIGENCE MODELS
Publication number
US20260350174A1
Application number
19171847
Filed
Apr 7, 2025
Published
Nov 5, 2026
Inventors
1
Claims
20

Classified under

G06F 21/577 Physics & computingG06F 2221/033 Physics & computing

More from Nvidia in Security & privacy

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.