Information Processing for AI Model Ethics and Access Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
AI models created by users may fail to comply with AI ethics, leading to potential AI-ethics-related problems when published globally, such as focusing on specific individuals instead of intended targets.
Innovation Solution
An information processing apparatus that evaluates AI models for compliance with AI ethics and determines a publication range by comparing evaluation accuracy with reference values, allowing publication to all users, a subset of users, or only the original creator based on the model's accuracy and ethical compliance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If AI models are published globally to all users, then model accessibility and utility are improved, but AI-ethics-related problems and harmful effects increase
Solution Approach 1:
The patent applies local quality by differentiating publication ranges based on evaluation results. Models are published to different user groups (all users, subset of users, or only original creator) depending on their ethical compliance score, thereby providing tailored access control that addresses harmful effects while maintaining accessibility for compliant models
Solution Approach 2:
The patent introduces an intermediary evaluation mechanism that assesses AI models before publication. This intermediary system (evaluation unit) acts as a mediator between model creators and users, determining publication ranges based on ethical compliance to prevent harmful effects while allowing accessible use of compliant models
2Reliability
If AI models are strictly controlled and limited publication range is enforced, then AI-ethics compliance is improved, but model accessibility and utility decrease
Solution Approach 1:
The patent applies parameter changes by using evaluation accuracy scores as a parameter to determine publication ranges. The system changes the accessibility parameter based on the ethical compliance parameter, allowing models with high compliance scores to be published broadly while limiting access to models with lower scores, thus balancing reliability and adaptability
3Reliability
If AI models are evaluated thoroughly for ethical compliance, then AI-ethics problems are reduced, but processing time and system complexity increase
Solution Approach 1:
The patent applies universality by designing an evaluation unit that can assess multiple AI models using the same ethical compliance framework. The evaluation system serves multiple functions (compliance assessment, publication range determination, risk evaluation) thereby reducing overall system complexity while maintaining thorough ethical checking
Data Source
AI summary
An information processing apparatus that can communicate with another information processing apparatus, the information processing apparatus comprising: an acquisition unit configured to acquire information about a trained AI model from the other information processing apparatus; an evaluation unit configured to evaluate the trained AI model; and a determination unit configured to, based on a result of the evaluation by the evaluation unit, determine a publication range within which the information processing apparatus will publish the trained AI model.