Real-time Privacy Filter for Caller Audio Masking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing communication systems fail to effectively prevent misuse of sensitive personal information (SPI) during live interactions between users and agents, despite the need for such information to complete transactions or authenticate identities, leading to potential harm.
Innovation Solution
A masking system acts as an intermediary between callers and agents, using automatic speech recognition and natural language processing to detect and redact SPI from audio streams, ensuring that only masked information is shared with agents while allowing SPI to be passed securely to organizational systems for transaction purposes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If SPI is collected from users to complete transactions and authenticate identities, then the organization can perform necessary business functions, but the risk of agent misuse of SPI increases
Solution Approach 1:
The patent extracts SPI from the audio stream transmitted to the agent by using automatic speech recognition to identify SPI in the caller's speech, then redacts it before transmission. This separates the SPI collection function from the agent's reception, allowing the organization to obtain necessary information while preventing agent access to sensitive data.
Solution Approach 2:
The patent introduces an intermediary system between the caller and agent that includes an automatic speech recognition component and a redaction component. This intermediary automatically processes the audio stream, identifies SPI, and selectively blocks it from reaching the agent while allowing legitimate transaction information to pass through.
2Object-affected harmful factors
If SPI is redacted from the audio stream to prevent agent access, then SPI misuse is prevented, but the agent cannot use SPI for legitimate transaction purposes
Solution Approach 1:
The system extracts only the specific SPI portions from the audio stream using automatic speech recognition and natural language processing, rather than redacting entire conversations. This allows the agent to hear and process legitimate transaction information while SPI is selectively removed.
Solution Approach 2:
The system provides feedback to the agent through the user interface when SPI is detected and redacted, allowing the agent to understand that sensitive information was present and take appropriate actions through the organization system to obtain necessary SPI for transaction completion.
3Reliability
If background checks and surveillance are used to prevent SPI misuse, then security monitoring is provided, but the complexity and cost of the system increase
Solution Approach 1:
The system uses automatic speech recognition and natural language processing to autonomously identify and redact SPI without requiring human agents to manually monitor or report sensitive information. The system self-regulates by automatically detecting SPI patterns and applying redaction rules based on organizational policies.
Solution Approach 2:
The patent replaces manual surveillance and background check processes with automated computational systems that use speech recognition and pattern matching to identify SPI. This substitution of mechanical/human processes with automated systems reduces operational complexity while maintaining or improving security effectiveness.
Data Source
AI summary
A masking system prevents a human agent from receiving sensitive personal information (SPI) provided by a caller during caller-agent communication. The masking system includes components for detecting the SPI, including automated speech recognition and natural language processing systems. When the caller communicates with the agent, e.g., via a phone call, the masking system processes the incoming caller audio. When the masking system detects SPI in the caller audio stream or when the masking system determines a high likelihood that incoming caller audio will include SPI, the caller audio is masked such that it cannot be heard by the agent. The masking system collects the SPI from the caller audio and sends it to the organization associated with the agent for processing the caller's request or transaction without giving the agent access to caller SPI.