Generative AI Abstraction for Privacy-Preserved Data Services
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for handling sensitive consumer data in electronic payment networks require significant computational resources and compliance measures, often without user consent, leading to inefficiencies and privacy concerns.
Innovation Solution
A system utilizing generative AI abstraction to create abstracted datasets from user profiles, allowing for privacy-preserved data services by generating user profiles based on identification data, associating these datasets with user profiles, and providing responses to third-party queries using these abstracted datasets instead of actual user data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If sensitive consumer data is collected and processed by transaction service providers to improve relationships and interactions, then service quality and personalization are improved, but computational resources and compliance costs increase significantly
Solution Approach 1:
The patent creates synthetic copies of consumer data through generative AI models that replicate the statistical properties and patterns of real data without containing actual sensitive information. These synthetic datasets enable service personalization and model training while eliminating the need to process and store real consumer data, significantly reducing computational resources and compliance costs.
Solution Approach 2:
The patent introduces synthetic data as an intermediary layer between real consumer data and AI models. Instead of directly processing sensitive consumer data, the system uses synthetic data generated by generative models to train and evaluate AI algorithms, thereby achieving service personalization without the computational burden and compliance requirements of handling real data.
2Adaptability or versatility
If consumer data is collected and stored to enable data services and queries, then system functionality and interoperability are improved, but data privacy risks and security requirements increase
Solution Approach 1:
The patent generates synthetic copies of consumer data that preserve the statistical properties, relationships, and patterns necessary for system interoperability and data services. These synthetic datasets enable third-party queries and analytics while containing no actual sensitive consumer information, thereby eliminating privacy risks while maintaining full system functionality.
Solution Approach 2:
The patent extracts only the essential statistical properties and patterns from real consumer data to create synthetic datasets. By separating the useful analytical characteristics from the sensitive personal information, the system enables interoperability and data services while completely removing privacy risks associated with storing and processing real consumer data.
3Measurement precision
If real consumer data is used for training machine learning models, then model accuracy and performance are improved, but user consent requirements and compliance measures increase
Solution Approach 1:
The patent creates synthetic training data that replicates the statistical properties, distributions, and relationships of real consumer data. These synthetic datasets enable machine learning models to learn the same patterns and achieve equivalent accuracy without requiring actual consumer data, thereby eliminating consent requirements and simplifying compliance measures while maintaining model performance.
Data Source
AI summary
Systems, methods, and computer program products are provided for privacy-preserved data services using generative AI abstraction. The system includes at least one processor configured to generate a user profile based on identification data of a user by inputting the identification data to a generative machine learning model, generate abstracted datasets based on outputs of the generative machine learning model, and associate abstracted datasets with the user profile. The at least one processor is further configured to receive a request message from a third-party computing device comprising a query and a token associated with the user profile, determine the user profile based on the token, and generate outputs based on the query and the abstracted datasets associated with the user profile. The at least one processor is further configured to communicate a response message to the third-party computing device based on the outputs.