Browser AI Prompt Screening to Prevent Sensitive Data Disclosure

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing systems and techniques disclosed herein address the challenge of balancing the access to generative AI functionalities, while minimizing the risk of compromising the confidentiality of sensitive information by preventing the transmission of such information to remote servers hosting generative AI tools, particularly when accessed through chat-style interfaces, in-line applications, or web browsers.

Innovation Solution

Implementing a computer-implemented method that includes comparing the URL of a webpage to configuration information stored on the endpoint device, scanning data entries for sensitive information, and using a custom handler to prevent transmission to remote servers if sensitive information is detected, while allowing transmission if no sensitive information is found.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If access to generative AI technologies is facilitated through web browsers, then productivity and ease of operation are improved, but the risk of sensitive information disclosure increases

Engineering Contradiction:
ImproveAI-assisted business productivityVSAvoidsensitive information disclosure
Core Design Contradiction:
ProductivityVSObject-affected harmful factors

Solution Approach 1:

The system performs preliminary actions by scanning data entries for sensitive information before transmission to the generative AI tool occurs. The scanning engine analyzes the data entry and custom handler intercepts the transmission process, preventing sensitive information from being sent to remote servers while allowing legitimate requests to proceed.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

A custom handler is introduced as an intermediary component between the web browser and the remote generative AI tool. This custom handler intercepts data entries, transfers them to the scanning engine for analysis, and controls whether the data is transmitted to the remote server based on the scanning results, thereby mediating the communication to prevent sensitive information disclosure.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If data entries are scanned for sensitive information before transmission, then security is improved, but device complexity and processing time increase

Engineering Contradiction:
Improveconfidentiality protectionVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The custom handler serves multiple functions: it intercepts data entries for scanning, transfers data to the scanning engine, and controls transmission to the remote server. This multi-functionality consolidates security operations into a single component, reducing overall system complexity while maintaining comprehensive protection.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The scanning engine automatically analyzes data entries for sensitive information without requiring manual intervention. The system self-regulates the transmission process by using the scanning results to automatically allow or block data transmission, reducing operational complexity while maintaining security.

Inventive Principle:
Principle #25Self-service

3Object-affected harmful factors

If transmission of data entries is blocked when sensitive information is detected, then confidentiality is improved, but ease of operation deteriorates

Engineering Contradiction:
Improvesensitive information disclosureVSAvoiduser interaction
Core Design Contradiction:
Object-affected harmful factorsVSEase of operation

Solution Approach 1:

The system provides feedback to users when data entries are blocked due to sensitive information detection. The custom handler and scanning engine work together to inform users about the blocking action, allowing them to understand why transmission was prevented and potentially modify their input, thereby maintaining ease of operation while ensuring security.

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS20260003980A1Preventing Disclosure of Sensitive Information In Attempts to Utilize Generative Artificial Intelligence (AI) Through a Web Browser Interface
Publication Date: 2026.01.01 PROOFPOINT INC
  • US20260003980A1 patent drawing
  • US20260003980A1 patent drawing
  • US20260003980A1 patent drawing

AI summary

A computer-implemented method is described for controlling disclosure of sensitive information to a remote network destination includes displaying a webpage on a graphical user interface of a computer with a web browser, receiving a data entry (e.g., a prompt for a generative artificial intelligence tool at the remote network destination) in a text box on the webpage at the computer, comparing the webpage's uniform resource locator to configuration information (e.g., that includes a list of web resources) stored in memory on the computer, and determining whether to scan the data entry for sensitive information (and potentially block transmission of the data entry) depending on whether the comparison identifies a match between the URL and the configuration information.