# US 2025/0159014 A1 · Blockchain enabled large languade model red teaming

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2025/0159014 A1
- Document ID: US20250159014A1
- Status: Published application
- Kind: A1
- Application number: 18/510,400
- Filed: 2023-11-15
- Published: 2025-05-15
- Earliest claimed: 2023-11-15
- Assignees: American Express Travel Related Services Company, Inc.
- Applicants: American Express Travel Related Services Company, Inc.
- Inventors: Hiranmayi Palanki; Shankar Djeyassilane
- Web page: https://patentgrep.com/patent/US20250159014A1

## CPC

- H04L 63/1433: {Vulnerability analysis}
- H04L 9/3239: {involving non-keyed hash functions, e.g. modification detection codes [MDCs], MD5, SHA or RIPEMD}
- H04L 9/50: {using hash chains, e.g. blockchains or hash trees}
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

Disclosed are various approaches for blockchain-enabled large language model (LLM) red teaming. A large language model (LLM) red teaming management agent executed in a distributed ledger environment can deploy a plurality of subsidiary agents in the distributed ledger environment. An LLM red teaming agent deployed in the distributed ledger environment can cause a network service to perform an LLM red teaming action and return an LLM red teaming result. The LLM red teaming result can be evaluated to generate endorsement status information for the LLM.

## Claims

### Claim 1

A system, comprising:
at least one computing device comprising at least one processor and at least one memory; and
machine-readable instructions stored in the at least one memory that, when executed by the at least one processor, cause the at least one computing device to at least:
execute, in a distributed ledger environment, a large language model (LLM) red teaming management agent that initiates a security analysis of an LLM by deploying a plurality of subsidiary agents in the distributed ledger environment, wherein the subsidiary agents perform actions for the security analysis of the LLM;
execute an LLM red teaming agent of the plurality of subsidiary agents, wherein the LLM red teaming agent transmits a request for a network service to perform an LLM red teaming action for the security analysis of the LLM, and the LLM red teaming agent receives an LLM red teaming result of the LLM red teaming action; and
execute an LLM evaluation agent of the plurality of subsidiary agents, wherein the LLM evaluation agent evaluates the LLM red teaming result to generate endorsement status information for the LLM.

### Claim 2 (depends on 1)

The system of claim 1, wherein the LLM red teaming management agent is deployed in public nodes of the distributed ledger environment, and the plurality of subsidiary agents are deployed in private nodes of the distributed ledger environment.

### Claim 3 (depends on 1)

The system of claim 1, wherein the LLM red teaming agent generates a graphical user interface or a programmatic interface that receives the LLM red teaming result from the network service.

### Claim 4 (depends on 1)

The system of claim 1, wherein the LLM evaluation agent transmits at least a portion of the endorsement status information to the LLM red teaming management agent.

### Claim 5 (depends on 1)

The system of claim 1, wherein the plurality of subsidiary agents further comprises at least one of: a software bill of materials (SBOM) agent that generates an SBOM based at least in part on the LLM, a trusted software supply chain (TSSC) agent that identifies a signature based at least in part on the LLM, or any combination thereof.

### Claim 6 (depends on 1)

The system of claim 1, wherein the plurality of subsidiary agents further comprises an LLM security agent that transmits instructions that enable or disable communications with the LLM based at least in part on the endorsement status information.

### Claim 7 (depends on 1)

The system of claim 1, wherein the LLM evaluation agent transmits a notification comprising at least a portion of the endorsement status information to an LLM network service that provides the LLM.

### Claim 8

A method, comprising:
initiating, by a large language model (LLM) red teaming management agent executed in a distributed ledger environment, a security analysis of an LLM by deploying a plurality of subsidiary agents in the distributed ledger environment;
transmitting, by an LLM red teaming agent deployed by the LLM red teaming management agent, a request for a network service to perform an LLM red teaming action, wherein the LLM red teaming agent receives an LLM red teaming result of the LLM red teaming action from the network service; and
evaluating, by an LLM evaluation agent deployed by the LLM red teaming management agent, the LLM red teaming result to generate endorsement status information for the LLM.

### Claim 9 (depends on 8)

The method of claim 8, further comprising:
transmitting, by the LLM evaluation agent, at least a portion of the endorsement status information to a publicly available network location.

### Claim 10 (depends on 8)

The method of claim 8, wherein the LLM red teaming agent and the LLM evaluation agent are executed using private nodes of the distributed ledger environment.

### Claim 11 (depends on 8)

The method of claim 8, further comprising:
transmitting, by an LLM security agent deployed by the LLM red teaming management agent, instructions that enable or disable communications with the LLM.

### Claim 12 (depends on 11)

The method of claim 11, wherein the LLM security agent is executed using at least one of: private nodes of the distributed ledger environment, sidechain nodes of the distributed ledger environment, or any combination thereof.

### Claim 13 (depends on 8)

The method of claim 8, further comprising:
generating, by the LLM red teaming agent, a graphical user interface or a programmatic interface; and
receiving the LLM red teaming result from the network service based at least in part on the graphical user interface or the programmatic interface.

### Claim 14 (depends on 8)

The method of claim 8, wherein the plurality of subsidiary agents comprise at least one of: a software bill of materials (SBOM) agent that generates an SBOM based at least in part on the LLM, a trusted software supply chain (TSSC) agent that identifies a signature based at least in part on the LLM, or any combination thereof.

### Claim 15

A system, comprising:
at least one computing device comprising at least one processor and at least one memory; and
machine-readable instructions stored in the at least one memory that, when executed by the at least one processor, cause the at least one computing device to at least:
deploy, by a large language model (LLM) red teaming management agent executed in a distributed ledger environment, a plurality of subsidiary agents in the distributed ledger environment;
transmit, by an LLM red teaming agent deployed by the LLM red teaming management agent, a notification for a network service to perform an LLM red teaming action and return, to the LLM red teaming agent, an LLM red teaming result of the LLM red teaming action; and
evaluate, by an LLM evaluation agent deployed by the LLM red teaming management agent, the LLM red teaming result to generate endorsement status information for the LLM.

### Claim 16 (depends on 15)

The system of claim 15, wherein the LLM red teaming action comprises at least one of: a red-teaming-based harmful content mitigation test of the LLM, a red-teaming-based bias mitigation test of the LLM, a red-teaming-based sensitive data element (SDE) leakage prevention test of the LLM, a red-teaming-based hallucination mitigation test of the LLM, a red-teaming-based LLM threat model test of the LLM, a red-teaming-based prompt injection security test of the LLM, or any combination thereof.

### Claim 17 (depends on 15)

The system of claim 15, wherein the LLM red teaming result comprises at least one of: a score generated by the LLM red teaming action, a document generated by the LLM red teaming action, a suggested corrective action generated by the LLM red teaming action, an example of an inappropriate interaction identified by the LLM red teaming action, an example of an inappropriate LLM response identified by the LLM red teaming action, or any combination thereof.

### Claim 18 (depends on 15)

The system of claim 15, wherein the machine-readable instructions, when executed by the at least one processor, further cause the at least one computing device to at least:
transmit, by an LLM security agent deployed by the LLM red teaming management agent, instructions that enable or disable communications with the LLM.

### Claim 19 (depends on 18)

The system of claim 18, wherein the LLM security agent is executed using at least one of: private nodes of the distributed ledger environment, sidechain nodes of the distributed ledger environment, or any combination thereof.

### Claim 20 (depends on 15)

The system of claim 15, wherein the machine-readable instructions, when executed by the at least one processor, further cause the at least one computing device to at least:
transmit, by the LLM evaluation agent, at least a portion of the endorsement status information to a publicly available network location.
