What does this research mean for the field?

AI Integrity can be quantitatively measured through Value Discovery Profiling and Entropy Analysis, revealing Integrity Hallucinations in large language models. Novelty: ClaimNovelty.NOVEL_FINDING. Consensus alignment: ConsensusAlignment.ESTABLISHES_NEW_DIRECTION.

What question did this study set out to answer?

The aim is to establish a measurable standard for AI's decision-making consistency through value profiling and entropy metrics.

February 16, 2026Open Access

AI Integrity Standard: A Quantitative Measurement Framework via Value Discovery Profiling and Entropy Analysis

Key Points

The aim is to establish a measurable standard for AI's decision-making consistency through value profiling and entropy metrics.
Developed a framework named Value Discovery Profiling utilizing 11,340 unique dilemma scenarios.
Analyzed decision-making using Value Entropy (H_V) and Integrity Gap (IG).
Provided empirical benchmarks for Llama 3.3, Claude 4.5, and Gemini 3.
Identified 'Integrity Hallucinations' where AI models produce conflicting ethical outputs.
Quantified AI decision-making consistency with defined metrics.

Abstract

This paper defines "AI Integrity" as a measurable standard for AI's decision-making consistency. We propose "Value Discovery Profiling, " a framework that utilizes 11, 340 unique dilemma scenarios to quantify how LLMs internalize human values. Through "Value Entropy (HV) " and the "Integrity Gap (IG), " we identify "Integrity Hallucinations" where models generate conflicting ethical logics. Empirical benchmarks of Llama 3. 3, Claude 4. 5, and Gemini 3 are provided.

Read Full Paperexternally

Mark Helpful

Bookmark

Relay

View Full Paper