Co-browsing Document Redaction via AI Authorization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In co-browsing sessions, users may inadvertently share sensitive information with unauthorized participants due to lack of control over document visibility, as they are unaware of the participants' identities and authorization levels, leading to unnecessary exposure of confidential data.
Innovation Solution
An AI agent dynamically analyzes documents and applies redactions based on participants' authorization levels, ensuring only authorized users can view sensitive information, with optional user override and training to improve accuracy over time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a user shares a document with all participants in a co-browsing session, then all participants can view the document, but unauthorized participants may view sensitive information
Solution Approach 1:
The system performs preliminary redaction of sensitive information from documents before presenting them to participants. The AI agent analyzes the document content, identifies sensitive portions, and applies appropriate redactions based on participant authorization levels before the document is viewed, preventing harmful information exposure in advance
Solution Approach 2:
Different portions of the document are treated differently based on their sensitivity and the participant's authorization level. The system applies selective redaction to specific portions of the document rather than uniformly blocking or allowing all content, ensuring each participant receives appropriately filtered information based on their role
2Object-affected harmful factors
If the user manually identifies and redacts portions of the document for each participant, then security is improved, but the complexity and time required increases
Solution Approach 1:
The system employs an AI agent that automatically analyzes document content, identifies sensitive portions, and applies appropriate redactions without requiring manual user intervention. The AI agent self-manages the security control process by comparing participant authorization levels with document sensitivity, eliminating the need for users to manually configure security settings
Solution Approach 2:
The system incorporates feedback mechanisms where user corrections of AI redaction decisions are captured and used to retrain the AI model. This feedback loop continuously improves the AI's accuracy in identifying sensitive information and applying appropriate redactions, reducing system complexity over time as the AI becomes more proficient
3Ease of operation
If the AI agent automatically redacts document portions, then security is enhanced and operation is simplified, but accuracy may initially be insufficient
Solution Approach 1:
The system captures user feedback when users correct or override AI redaction decisions. This feedback is used to retrain the AI model, continuously improving its accuracy in identifying sensitive information. The feedback mechanism transforms initial imprecision into improved performance over time through iterative learning
Solution Approach 2:
The system replaces manual human judgment with an AI-based automated system that uses machine learning to identify and redact sensitive information. While the AI initially lacks the precision of human experts, it provides consistent automated operation that can be continuously improved through training, eventually achieving or exceeding human-level accuracy
Data Source
AI summary
Co-browsing allows a providing party to access visual content on a computing device for sharing with one or more other parties. The parties receiving the shared image may have dissimilar security authorizations. Accordingly, systems and methods are provided that enable shared content, such as a document, web page viewed in a browser, etc., to automatically be redacted to block those parties who are not authorized to view the content. For example, a neural network may be utilized to scan the document and provide specific redacted copies to the parties so each can view the image of the content with unauthorized content redacted.