PII Tokenization in Audio Video Recordings for Access Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems fail to provide universal employee access to digital audio and video recordings of customer information while maintaining compliance and protecting sensitive personal or private information, as employees without proper clearance cannot view or use recordings containing such information without risking confidentiality or security breaches.
Innovation Solution
A system that uses a machine learning model to identify personally identifiable information (PII) in digital media content and tokenizes it based on user access levels, allowing authorized employees to access the content while keeping sensitive information hidden from those without the necessary clearance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If employees are granted universal access to digital audio and video recordings containing customer information, then productivity and ease of operation improve, but information security and confidentiality are compromised for unauthorized personnel
Solution Approach 1:
The system segments customer information into different sensitivity levels (PII, sensitive, non-sensitive) and applies different access controls to each segment. Employees can access recordings based on their clearance level, with automated redaction or masking of information segments they are not authorized to view. This allows universal access to recordings while protecting sensitive segments from unauthorized viewing.
Solution Approach 2:
An automated access control system acts as an intermediary between employees and customer information in recordings. The system automatically analyzes recordings, identifies sensitive information, and applies appropriate redaction or masking based on employee clearance levels. This intermediary layer enables employees to access recordings without direct exposure to protected information segments.
2Reliability
If sensitive customer information is protected through restricted access, then information security improves, but productivity and ease of operation deteriorate due to limited employee access
Solution Approach 1:
The system dynamically adjusts access controls based on employee clearance levels and the sensitivity of information in each recording. Access permissions are not static but adapt automatically - employees with higher clearances can access more sensitive information, while those with lower clearances have automatically redacted versions. This dynamic approach maintains strong protection for sensitive data while allowing efficient access for authorized personnel.
Solution Approach 2:
The automated access control system performs self-service by automatically analyzing recordings, identifying sensitive information, determining appropriate clearance levels, and applying redaction or masking without manual intervention. This eliminates the need for manual review and approval processes, maintaining strong security protections while enabling employees to efficiently access recordings within their authorization levels.
3Reliability
If automated redaction or masking is applied to protect PII, then information security improves, but manufacturing precision and measurement precision worsen due to potential over-redaction or failure to detect all PII
Solution Approach 1:
The system incorporates feedback mechanisms where redaction decisions are continuously refined based on analysis of redacted content, employee access patterns, and security outcomes. The automated access control system learns from each access event and adjusts its redaction patterns to improve accuracy over time, reducing both over-redaction and under-redaction while maintaining strong PII protection.
Data Source
AI summary
Various embodiments are directed to a system for identifying personally identifiable information (PII) in digital media content, such as audio files, videos, images, etc. and providing such content with one or more portions thereof appropriately tokenized based on an access level of the user requesting the content. The PII may be detected in the digital media content using a machine learning model or a classification model.