ArticleNature computational science2023
Human-like intuitive behavior and reasoning biases emerged in large language models but disappeared in ChatGPT.
Article in Nature computational science, 2023. The graph could read no effect estimate from its abstract, so it casts no vote on the map. Cited by 38 papers.
What it found
Each row is one number read from the abstract, on the scale the paper reported it, with its interval. Left of the dashed line favours the treatment, right favours the comparator. Under each row is the sentence it came from. New to these charts? A ten-minute tutorial.
The abstract states no effect estimate the extractor could read, or names no intervention and outcome on the map, so this paper lights no cell and moves no belief. It is still indexed, cited and linked below.
The trial behind it
Trials whose registry record cites this paper, or whose number appears in the abstract. A trial that started after this paper was published is citing it as background, not reporting it.
Neither the registry nor the abstract names a trial number. If this is a trial report, that itself is worth knowing.
Who cites it
38 citing papers in PubMed, 38 citations in OpenAlex.
- The Internalized Ableism Inventory: Scale development using a hybrid artificial intelligence and community-based participatory research design.Rehabilitation psychology · 2026Article
- Understanding large language models demands distinguishing human projection from machine cognition.Communications psychology · 2026Review
- WiNGPT-32B: An Open-Source, Locally Deployable LLM for RECIST Assessment via Chained Task Execution Using Radiology Report Text.Diagnostics (Basel, Switzerland) · 2026Article
- AIPsychoBench: Understanding the Psychometric Differences Between LLMs and Humans.Topics in cognitive science · 2026Article
- LLMs displaying less cognitive bias are not necessarily better decision makers.Nature machine intelligence · 2026Article
- Toward a science of human-AI teaming for decision making: A complementarity framework.PNAS nexus · 2026Article
- Quantifying uncert-AI-nty: Testing the accuracy of LLMs' confidence judgments.Memory & cognition · 2026Article
- Influence of solution efficiency and valence of instruction on additive and subtractive solution strategies in humans, GPT-4, and GPT-4o.Communications psychology · 2026Article
- AMULED: Addressing Moral Uncertainty using Large language models for Ethical Decision-making.Frontiers in artificial intelligence · 2026Article
- A moral Turing test: How belief and source shape detection of and agreement with LLM judgments.PloS one · 2026Article
- Cross-linguistic patterns of cognitive biases in large language models: a comparative study in English, Hebrew, and Russian.Frontiers in artificial intelligence · 2026Article
- Divergent patterns of probabilistic reasoning in humans and GPT-5.Frontiers in psychology · 2026Article
- Leveraging imperfection with MEDLEY: a multi-model approach harnessing bias in medical AI.Frontiers in artificial intelligence · 2026Article
- Revealing emergent human-like conceptual representations from language prediction.Proceedings of the National Academy of Sciences of the United States of America · 2025Article
- ChatGPT as therapy: A qualitative and network-based thematic profiling of shared experiences, attitudes, and beliefs on Reddit.Journal of psychiatric research · 2025Article
- Large language models for autism: evaluating theory of mind tasks in a gamified environment.Scientific reports · 2025Article
- Humans and LLMs rate deliberation as superior to intuition on complex reasoning tasks.Communications psychology · 2025Article
- Evaluating the ability of large Language models to predict human social decisions.Scientific reports · 2025Article
- Pitfalls of large language models in medical ethics reasoning.NPJ digital medicine · 2025Article
- Large language models show amplified cognitive biases in moral decision-making.Proceedings of the National Academy of Sciences of the United States of America · 2025Article
Corrections and comments
PubMed lists nothing against this paper. Absence here is not a guarantee, only a check that was made.
Authors and funding
3 authors at 2 institutions in 2 countries.
Funding
Abstract
We design a battery of semantic illusions and cognitive reflection tests, aimed to elicit intuitive yet erroneous responses. We administer these tasks, traditionally used to study reasoning and decision-making in humans, to OpenAI's generative pre-trained transformer model family. The results show that as the models expand in size and linguistic proficiency they increasingly display human-like intuitive system 1 thinking and associated cognitive errors. This pattern shifts notably with the introduction of ChatGPT models, which tend to respond correctly, avoiding the traps embedded in the tasks. Both ChatGPT-3.5 and 4 utilize the input-output context window to engage in chain-of-thought reasoning, reminiscent of how people use notepads to support their system 2 thinking. Yet, they remain accurate even when prevented from engaging in chain-of-thought reasoning, indicating that their system-1-like next-word generation processes are more accurate than those of older models. Our findings highlight the value of applying psychological methodologies to study large language models, as this can uncover previously undetected emergent characteristics.
Indexed as
Identifiers
What OpenQuestion holds
Registered trials
Read under generation 80e0d062 · epoch 390. Bibliography from PubMed, PubMed Central and OpenAlex; grants from NIH RePORTER; trial links from ClinicalTrials.gov; estimates, votes and beliefs from the OpenQuestion graph.