AI Ethical Risk Evaluation Using Constraint-Based Rule Prioritization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing AI systems face challenges in identifying high-priority ethical risk check items due to meaningless adoption of rules based on past incident cases, particularly when limitation conditions are applied, leading to overlooked ethical risks.

Innovation Solution

An evaluation program that identifies sets of constituents related to an AI system, selects rules based on limitation conditions, and determines priorities to output an AI ethical risk evaluation result, thereby extracting high-priority check items.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If rules based on past incident cases are adopted without considering limitation conditions, then the evaluation process is simplified, but high-priority check items may be overlooked and ethical risk assessment accuracy deteriorates

Engineering Contradiction:
Improveevaluation process complexityVSAvoidethical risk assessment accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent implements dynamic rule selection by determining whether limitation conditions exist for each constituent before applying rules. The evaluation process adapts its behavior based on the presence or absence of limitation conditions, switching between different rule application modes to maintain both efficiency and accuracy.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent applies different evaluation strategies to different parts of the system based on local characteristics. When limitation conditions are present, a specialized rule determination process is applied; when absent, the standard rule-based approach is used. This localized adaptation ensures high-priority check items are not overlooked while maintaining overall process efficiency.

Inventive Principle:
Principle #3Local quality

2Reliability

If all rules from past incident cases are applied uniformly, then comprehensive coverage is achieved, but meaningless check items increase evaluation time and reduce efficiency

Engineering Contradiction:
Improvecheck item coverageVSAvoidevaluation efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent extracts and removes meaningless check items by determining whether limitation conditions exist for each constituent. Rules that would generate meaningless check items under certain limitation conditions are excluded from application, thereby improving evaluation efficiency without sacrificing coverage of meaningful ethical risks.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent dynamically adjusts the set of applicable rules based on the specific context of each constituent and its limitation conditions. This dynamic adaptation allows the system to maintain comprehensive coverage of relevant ethical risks while excluding meaningless check items, thereby improving evaluation efficiency.

Inventive Principle:
Principle #15Dynamics

3Measurement precision

If limitation conditions are considered in rule selection, then meaningful check items are identified, but the evaluation process complexity increases

Engineering Contradiction:
Improvecheck item relevanceVSAvoidrule selection process complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the rule selection process into distinct phases: first determining whether limitation conditions exist for each constituent, then selectively applying rules based on this determination. This segmentation simplifies the overall complexity by breaking down the decision-making process into manageable, conditional steps.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements a dynamic rule selection mechanism that adapts to the presence or absence of limitation conditions. This dynamic approach maintains check item relevance by selectively applying rules only when appropriate, while managing process complexity through structured conditional logic rather than exhaustive analysis.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4693120A1Evaluation program, evaluation method, and information processing device
Publication Date: 2026.02.11 FUJITSU LTD
  • EP4693120A1 patent drawingFigure 1
  • EP4693120A1 patent drawingFigure 2
  • EP4693120A1 patent drawingFigure 3

AI summary

An evaluation program causing a computer to execute processes of identifying a plurality of sets of constituents that are related to each other on the basis of configuration information including stakeholders of an artificial intelligence (AI) system, selecting one or more rules from among a plurality of rules on the basis of a limitation condition for the constituent, determining priorities of the plurality of specified sets on the basis of the one or more selected rules, and outputting an AI ethical risk evaluation result of the AI system on the basis of the determined priorities.