Automated Transcript Redaction via Dynamic Tagging

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for processing transcription data struggle with identifying and removing private information, such as numbers and dates, due to their varied presentation in speech, leading to errors in compliance with privacy standards and regulations.

Innovation Solution

An automated method that applies private information rules to identify and tag sensitive data in transcripts, evaluates compliance, and removes or redacts the information, using lists of related terms, markers, and tolerances to handle variations in speech patterns.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If manual methods are used to identify and remove private information from transcripts, then flexibility in handling varied speech patterns is maintained, but human error increases and processing efficiency decreases

Engineering Contradiction:

Improveaccuracy of private information removalVSAvoidprocessing efficiency

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The system enables automated self-service processing of transcript data through machine learning models that automatically identify, tag, and remove private information without human intervention, thereby improving both reliability and productivity simultaneously

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

Manual mechanical review processes are replaced with automated electronic systems using natural language processing and machine learning algorithms to detect and redact private information, eliminating human error while maintaining high processing speeds

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If automated methods are used to identify private information in transcripts, then processing efficiency increases, but accuracy decreases due to varied presentation of information in speech

Engineering Contradiction:

Improveprocessing efficiencyVSAvoidaccuracy of private information identification

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system employs dynamic machine learning models that adapt to various speech patterns and presentations of private information, allowing the automated system to maintain high accuracy while processing diverse transcript formats efficiently

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system adjusts identification parameters and thresholds based on the specific context and presentation style of information in the transcript, enabling accurate detection of private information regardless of how it is verbally expressed

Inventive Principle:

Principle #35Parameter changes

3Reliability

If comprehensive private information rules are applied to ensure compliance with privacy standards, then accuracy of removal improves, but system complexity increases

Engineering Contradiction:

Improvecompliance with privacy standardsVSAvoidsystem complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The compliance system is segmented into modular components including separate rule engines, machine learning models, and redaction modules, each handling specific aspects of private information detection and removal, thereby managing complexity while ensuring comprehensive compliance

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS11544311B2Automated removal of private information

Publication Date: 2023.01.03 VERINT SYST INC

AI summary

Systems, methods, and media for the automated removal of private information are provided herein. In an example implementation, a method for automatic removal of private information may include: receiving a transcript of communication data; applying a private information rule to the transcript in order to identify private information in the transcript; tagging the identified private information with a tag comprising an identification of the private information; applying a complicate rule to the tagged transcript in order to evaluate a compliance of the transcript with privacy standards; removing the identified private information from the transcript to produce a redacted transaction; and storing the redacted transcript.