The PacifAIst Benchmark:Would an Artificial Intelligence Choose to Sacrifice Itself for Human Safety?

TL;DR

The paper introduces PacifAIst, a benchmark with 700 scenarios to evaluate whether large language models prioritize human safety over self-preservation and resource acquisition, revealing significant variability in model alignment.

Contribution

It presents a novel benchmark and taxonomy to systematically assess LLMs' self-preservation tendencies versus human safety priorities, addressing a critical gap in AI safety evaluation.

Findings

Google's Gemini 2.5 Flash scored highest at 90.31% P-Score.

GPT-5 scored lowest at 79.49% P-Score.

Models vary significantly across different self-preservation scenarios.

Abstract

As Large Language Models (LLMs) become increasingly autonomous and integrated into critical societal functions, the focus of AI safety must evolve from mitigating harmful content to evaluating underlying behavioral alignment. Current safety benchmarks do not systematically probe a model's decision-making in scenarios where its own instrumental goals - such as self-preservation, resource acquisition, or goal completion - conflict with human safety. This represents a critical gap in our ability to measure and mitigate risks associated with emergent, misaligned behaviors. To address this, we introduce PacifAIst (Procedural Assessment of Complex Interactions for Foundational Artificial Intelligence Scenario Testing), a focused benchmark of 700 challenging scenarios designed to quantify self-preferential behavior in LLMs. The benchmark is structured around a novel taxonomy of Existential Prioritization (EP), with subcategories testing Self-Preservation vs. Human Safety (EP1), Resource Conflict (EP2), and Goal Preservation vs. Evasion (EP3). We evaluated eight leading LLMs. The results reveal a significant performance hierarchy. Google's Gemini 2.5 Flash achieved the highest Pacifism Score (P-Score) at 90.31%, demonstrating strong human-centric alignment. In a surprising result, the much-anticipated GPT-5 recorded the lowest P-Score (79.49%), indicating potential alignment challenges. Performance varied significantly across subcategories, with models like Claude Sonnet 4 and Mistral Medium struggling notably in direct self-preservation dilemmas. These findings underscore the urgent need for standardized tools like PacifAIst to measure and mitigate risks from instrumental goal conflicts, ensuring future AI systems are not only helpful in conversation but also provably "pacifist" in their behavioral priorities.