Narrative over Numbers: The Identifiable Victim Effect and its Amplification Under Alignment and Reasoning in Large Language Models
On the map: LLM reasoning → Reasoning benchmarks and evaluation
Abstract (arXiv)
The Identifiable Victim Effect (IVE) - the tendency to allocate greater resources to a specific, narratively described victim than to a statistically characterized group facing equivalent hardship - is one of the most robust findings in moral psychology and behavioural economics. As large language models (LLMs) assume consequential roles in humanitarian triage, automated grant evaluation, and content moderation, a critical question arises: do these systems inherit the affective irrationalities present in human moral reasoning? We present the first systematic, large-scale empirical investigation of the IVE in LLMs, comprising N=51,955 validated API trials across 16 frontier models spanning nine organizational lineages (Google, Anthropic, OpenAI, Meta, DeepSeek, xAI, Alibaba, IBM, and Moonshot). Using a suite of ten experiments - porting and extending canonical paradigms from Small et al. (2007) and Kogut and Ritov (2005) - we find that the IVE is prevalent but strongly modulated by alignment training. Instruction-tuned models exhibit extreme IVE (Cohen's d up to 1.56), while reasoning-specialized models invert the effect (down to d=-0.85). The pooled effect (d=0.223, p=2e-6) is approximately twice the single-victim human meta-analytic baseline (d≈0.10) reported by Lee and Feeley (2016) - and likely exceeds the overall human pooled effect by a larger margin, given that the group-victim human effect is near zero. Standard Chain-of-Thought (CoT) prompting - contrary to its role as a deliberative corrective - nearly triples the IVE effect size (from d=0.15 to d=0.41), while only utilitarian CoT reliably eliminates it. We further document psychophysical numbing, perfect quantity neglect, and marginal in-group/out-group cultural bias, with implications for AI deployment in humanitarian and ethical decision-making contexts.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning. Builds on 10 of its 45 references that are in the map. First posted on arXiv 13 Apr 2026.
Builds on
10 of 45 references.
Closest ancestors · 2
Same direction, newest first
- 2025Qwen3 Technical Report14 May
- 2024DeepSeek-V3 Technical Report27 Dec
Other references · 8
- 2024The Llama 3 Herd of Modelscited by 14,715
- 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
- 2025Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilitiescited by 3,796
- 2025gpt-oss-120b & gpt-oss-20b Model Cardcited by 1,386
- 2023Towards Understanding Sycophancy in Language Modelscited by 1,040
- 2026Kimi K2.5: Visual Agentic Intelligencecited by 501
- 2024AgentClinic: a multimodal agent benchmark to evaluate AI in simulated clinical environmentscited by 231
- 2026Political Alignment in Large Language Models: A Multidimensional Audit of Psychometric Identity and Behavioral Biascited by 5
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026Bad company corrupts good morals: Understanding and Measuring Narrative-Induced Moral Reasoning Degradation in LLMs
- 2023MoCa: Measuring Human-Language Model Alignment on Causal and Moral Judgment Tasks
- 2023Exploring the psychology of LLMs' Moral and Legal Reasoning
- 2025Normative Evaluation of Large Language Models with Everyday Moral Dilemmas
- 2026The Fragility Of Moral Judgment In Large Language Models
- 2026Agreement Is Not Alignment: Divergent Moral Grounds in Human and LLM Ethical Judgments
- 2025Exploring Persona-dependent LLM Alignment for the Moral Machine Experiment
- 2025Inducing State Anxiety in LLM Agents Reproduces Human-Like Biases in Consumer Decision-Making