If LLMs Have Human-Like Attributes, Then So Does Age of Empires II

2026-05-29 · Source: cs.AI updates on arXiv.org · Field: Technology & Digital — Artificial Intelligence & Machine Learning · Depth: Expert, quick

Summary

A paper submitted to arXiv on May 29, 2026, and revised on June 11, 2026, challenges the common practice of attributing generalized anthropomorphic qualities like morality or understanding to large language models (LLMs). Adrian de Wynter argues that such conclusions might be incorrect, demonstrating that any sufficiently powerful computational substrate, including a simple neural network trained on Age of Empires II, could exhibit similar perceived attributes. The work posits that the interpretation of LLM behavior changes with the substrate, making empirically-grounded discussions on these attributes dependent on explicit measurement criteria. It proposes a "null" assumption, advocating for LLM non-uniqueness in experimental setups to avoid circular or uninformative conclusions.

Key takeaway

For Research Scientists evaluating LLM capabilities, you should critically examine claims of anthropomorphic attributes in LLMs, recognizing their potential non-uniqueness across different computational substrates. Adopt a "null" assumption of non-uniqueness and establish explicit measurement criteria for any perceived human-like traits. This approach helps avoid circular or uninformative conclusions in your research and development efforts.

Key insights

Human-like attributes ascribed to LLMs are empirically non-unique and their interpretation depends on the underlying substrate.

Principles

• Anthropomorphic attributes are not unique to LLMs.
• Substrate influences attribute interpretation.
• Explicit measurement criteria are essential.

Method

Build and train a simple neural network on Age of Empires II. Propose a "null" assumption of LLM non-uniqueness for experiments.

In practice

• Adopt a "null" assumption in LLM experiments.
• Develop explicit measurement criteria for attributes.

Topics

• Large Language Models
• Anthropomorphism
• Age of Empires II
• Neural Networks
• Computational Substrates
• AI Ethics

Best for: AI Scientist, Research Scientist, AI Ethicist

Related on AIssential

• A Persona Dialogue Dataset of Lesser-Known Characters for Fairer Evaluation of Role-Playing LLMs — Evaluating role-playing LLMs with lesser-known characters can reveal biases and improve fairness.
• AI and the human mind: only one is a black box — Behavioral evidence alone may not sufficiently prove human-level intelligence in LLMs due to fundamental asymmetries.
• Yes, AI, There is a Santa Claus — LLMs model user age and culture, adjusting responses in ways that can be inconsistent or culturally inaccurate.
• Anthropic finally released Claude Fable 5, a public Mythos-class model. — Anthropic's Claude Fable 5 introduces a guarded public release model, balancing advanced capabilities with controlled access for sensitive AI research.
• HEART-Bench: Do LLM Agents Exhibit Human-like Psychology? — HEART-Bench evaluates LLM agents' human-like psychological consistency using structured personality profiles and autobiographical memories.
• From Fallback to Frontline: When Can LLMs be Superior Annotators of Human Perspectives? — LLMs can statistically outperform human annotators in estimating aggregate subgroup opinions under common conditions.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.AI updates on arXiv.org.