Selective Privacy Enforcement for Voice Application Data

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current voice application devices pose significant information privacy and confidentiality challenges due to their ability to record and store sensitive information without user awareness, especially when operating continuously, and existing methods for protecting such information often render entire passages un retrievable, making it difficult to balance privacy with the need for later access and analysis.

Innovation Solution

A computer-implemented method that identifies sensitive information in conversations, generates confidence scores, and determines appropriate protection actions based on these scores to form a modified conversation devoid of sensitive information, allowing for selective protection and secure storage while enabling later retrieval and analysis.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If sensitive information is protected by removing or redacting it from conversations, then privacy and confidentiality are improved, but the ability to retrieve and analyze the original information is lost

Engineering Contradiction:
Improveprivacy protectionVSAvoidinformation accessibility
Core Design Contradiction:
ReliabilityVSLoss of information

Solution Approach 1:

The patent segments the conversation into multiple parts: original conversation data, identified sensitive information portions, and redacted versions. This segmentation allows the system to preserve the complete original conversation for future analysis while simultaneously providing protected versions for privacy-sensitive contexts, thus resolving the contradiction between privacy protection and information accessibility

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary component (the privacy-preserving program) that acts as a mediator between the raw conversation data and its various uses. This intermediary identifies, marks, and manages sensitive information without destroying the original data, enabling both privacy protection and future retrieval/analysis of the complete conversation

Inventive Principle:
Principle #24Intermediary (Mediator)

2Productivity

If voice application devices continuously record conversations, then the ability to respond to user needs is improved, but the collection of sensitive information without user awareness increases

Engineering Contradiction:
Improveresponse capabilityVSAvoidunauthorized sensitive information collection
Core Design Contradiction:
ProductivityVSObject-generated harmful factors

Solution Approach 1:

The patent applies preliminary action by implementing automated identification and marking of sensitive information as it is being recorded during continuous conversation monitoring. The privacy-preserving program proactively detects and flags sensitive portions before they can be inadvertently misused or exposed, allowing the device to maintain continuous recording capability while preemptively protecting sensitive data

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system implements feedback mechanisms where the identified sensitive information portions are marked and tracked, providing continuous feedback about what sensitive data has been collected. This feedback loop enables the device to maintain productivity through continuous recording while being aware of and protecting sensitive information through automated identification and marking

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS11113419B2Selective enforcement of privacy and confidentiality for optimization of voice applications
Publication Date: 2021.09.07 INTERNATIONAL BUSINESS MACHINE CORPORATION
  • US11113419B2 patent drawing
  • US11113419B2 patent drawing
  • US11113419B2 patent drawing

AI summary

A computer-implemented method includes identifying a plurality of protected pieces from a conversation. The computer-implemented method further includes generating one or more confidence scores for each protected piece, wherein a confidence score is a degree of associativity between a protected piece and a type of sensitive information. The computer-implemented method further includes determining that the protected piece is associated with the type of sensitive information. The computer-implemented method further includes determining a type of protection action for each protected piece in the plurality of protected pieces. The computer-implemented method further includes performing the type of protection action for each protected piece in the plurality of protected pieces to form a modified conversation that is devoid of the sensitive information. A corresponding computer system and computer program product are also disclosed.