User-Type Ethicality Evaluation for AI Consultation Models

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing AI-based psychological consultation technologies lack methods for evaluating the ethicality of their output data, which is crucial for providing user-friendly responses and explanations tailored to user characteristics.

Innovation Solution

An electronic device with a processor that measures the ethicality degree of AI model output based on user types, using evaluation criteria stored in memory, and classifies users and AI models based on disposition, virtue, personality, cognitive faculty, and personal environments, while assessing interpretability, transparency, responsibility, bias, and stability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If AI-based psychological consultation is implemented as a black box system, then consultation accessibility and cost-effectiveness are improved, but ethicality evaluation capability deteriorates

Engineering Contradiction:
Improveconsultation accessibilityVSAvoidethicality evaluation capability
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The patent introduces an intermediary ethicality evaluation module that acts as a mediator between the AI consultation system and users. This module evaluates the ethicality of AI responses without requiring users to understand the internal black box mechanisms, thus maintaining accessibility while adding ethical assessment capability.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system is segmented into distinct functional modules: the AI consultation engine (black box), the ethicality evaluation module (separate assessment system), and the user interface. This segmentation allows the ethicality evaluation to operate independently while preserving the accessibility benefits of the AI system.

Inventive Principle:
Principle #1Segmentation

2Adaptability or versatility

If user type classification is added to evaluate ethicality, then user-customized consultation quality is improved, but system complexity increases

Engineering Contradiction:
Improveuser-customized consultation qualityVSAvoidsystem complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system dynamically adapts to different user types by adjusting evaluation criteria and weighting based on classified user characteristics. The ethicality evaluation parameters are not fixed but dynamically modified according to the user's psychological state, cognitive level, and consultation needs, enabling customization without requiring a completely separate system for each user type.

Inventive Principle:
Principle #15Dynamics

3Measurement precision

If multiple ethicality evaluation criteria are implemented, then ethicality measurement precision is improved, but evaluation process complexity increases

Engineering Contradiction:
Improveethicality measurement precisionVSAvoidevaluation process complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies different evaluation criteria and weighting schemes to different user types and consultation contexts. Rather than uniformly applying all ethicality criteria to all cases, the system selectively emphasizes relevant criteria based on local conditions such as user vulnerability, consultation topic sensitivity, and AI response type, thus improving measurement precision while managing complexity through contextual adaptation.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP4625430A1Apparatus for evaluating ethicality of artificial intelligence model according to user type
Publication Date: 2025.10.01 MOON MANKI
  • EP4625430A1 patent drawingFigure 1
  • EP4625430A1 patent drawingFigure 2
  • EP4625430A1 patent drawingFigure 3

AI summary

An electronic device according to an embodiment of the present invention may be configured by including: a memory for storing an evaluation criterion related to ethicality of an artificial intelligence model performing a consultation; and a processor that measures an ethicality degree of result data output by the artificial intelligence model according to the evaluation criterion, wherein the ethicality degree is measured according to types of users performing consultations with the artificial intelligence model.