LLMs Show Deep Similarities to Human Cognition

Chandra Sripada, Richard Lewis· July 31, 2026 View original

Key takeaways

  • LLMs exhibit deep cognitive similarities to human intelligence across five dimensions.
  • These parallels challenge the view of LLMs as fundamentally alien intelligences.
  • The findings suggest a broader model for understanding intelligent cognition.
  • Recognizing these similarities can inform better AI design and human-AI interaction.

Who benefits

AI ResearchSoftware DevelopmentEdTechConsultingPsychology

Summary

Researchers argue that large language models (LLMs) exhibit profound structural and functional similarities to human cognition across five key dimensions. This perspective challenges the view of LLMs as alien intelligences and suggests a broader model for understanding intelligence.

A recent paper posits that large language models (LLMs), despite their distinct physical substrates and learning histories, share deep cognitive organizational principles with human intelligence. This argument counters the prevailing notion that LLMs operate fundamentally differently from human minds. The authors identify five specific dimensions of convergence: inferential organization, computational architecture, representational structure, prediction-driven learning, and mechanisms akin to reinforcement learning for goal-directed actions. These correspondences are considered striking precisely because of the inherent differences between LLMs and humans. By highlighting these structural and functional parallels, the research advocates for a more inclusive model of intelligent cognition. This model would integrate core principles traditionally used to explain human intelligence with the observed characteristics of contemporary LLM-based systems, fostering a deeper understanding of both artificial and natural intelligence.

Why it matters

Understanding the cognitive parallels between LLMs and humans can inform better AI design, improve human-AI collaboration, and help professionals anticipate LLM behaviors. It also provides a framework for discussing the nature of intelligence itself, impacting strategic AI development.

How to implement this in your domain

  1. 1Consider these cognitive parallels when designing user interfaces and interaction patterns for LLM-powered applications.
  2. 2Leverage insights from human cognitive science to develop more intuitive and effective prompt engineering strategies.
  3. 3Explore how LLM 'cognitive' biases might mirror human biases and develop mitigation strategies.
  4. 4Use this framework to guide research into explainable AI, seeking to understand LLM reasoning in human-interpretable terms.
  5. 5Inform strategic discussions on the long-term trajectory of AI development, considering the implications of cognitive convergence.

Original post by Chandra Sripada, Richard Lewis

"arXiv:2607.26179v1 Announce Type: cross Abstract: LLMs are widely regarded as alien intelligences, systems whose cognitive operations are fundamentally unlike our own. Apparent similarities to human cognition are therefore often seen as the result of anthropomorphic projection. W…"

View on X

Originally posted by Chandra Sripada, Richard Lewis on X · view source

Want to go deeper?

Turn these trends into skills with Learnijoy's hands-on AI & tech courses.

Explore courses

More in AI Research

AI ResearchAI Engineering & DevTools

New Framework Improves Partial Multi-View Clustering Performance.

DAS-PMVC is a novel framework for partial multi-view clustering that addresses view asymmetry and irrelevant samples by leveraging dual alignment and structure enhancement. It uses anchor graph structure alignment, structure-enhanced feature learning, and a dual alignment strategy to achieve superior clustering performance on various datasets.

Shubin Ma, Liang Zhao, Chuanye He, Zhenjiao Liu, Liang Zou, Lin Yuanbo Wu, Yu ShaoJul 31, 2026
AI Engineering & DevToolsAI Research

Dual Teachers Improve Adversarial Robustness and Accuracy.

This work extends Information Bottleneck Distillation (IBD) by introducing a "clean teacher" alongside a robust teacher to improve the robustness/accuracy tradeoff against adversarial attacks. The proposed method transfers features from both teachers to a student model, achieving better clean accuracy while maintaining adversarial robustness, outperforming original IBD and competing with state-of-the-art approaches.

Vincent Ryusuke Takahashi, Yoshinari Takeishi, Jun'ichi Takeuchi, Kave SalamatianJul 31, 2026
AI Engineering & DevToolsAI Research

Dynamic Batch Sizes Improve Large Language Model Training Efficiency.

This paper proposes a new approach to deep learning dynamics, deriving joint scaling laws for loss based on both learning rate and batch size schedules. It introduces an optimal dynamic batch size schedule that consistently outperforms static batch size baselines, highlighting its importance for large language model training.

Jiaxiang Li, Zhiqi Bu, Shiyun XuJul 31, 2026