Related Experiment Video
Updated: Jul 4, 2025

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
Published on: December 6, 2024
Do large language models show decision heuristics similar to humans? A case study using GPT-3.5
Gaurav Suri1, Lily R Slater1, Ali Ziaee1
1Department of Psychology, San Francisco State University.
Large Language Models (LLMs) like ChatGPT exhibit cognitive biases, such as anchoring and framing effects, similar to humans. This suggests language itself may influence these context-sensitive responses.
Area of Science:
- Cognitive Science
- Artificial Intelligence
- Psychology
Background:
- Large Language Models (LLMs) are AI systems trained on extensive data to generate human-like text.
- Generative Pre-Trained Transformer (GPT)-3.5, powering ChatGPT, is a prominent example of an LLM.
- Human decision-making is known to be influenced by cognitive heuristics and context-sensitive responses.
Purpose of the Study:
- To investigate whether ChatGPT exhibits heuristics and context-sensitive responses.
- To compare ChatGPT's responses to those of human participants using identical prompts.
- To explore the role of language in cognitive biases.
Main Methods:
- Four studies were conducted using novel prompts to test ChatGPT and human participants.
- Study 1 examined the anchoring effect using random numerical anchors.
- Study 2 assessed representativeness and availability heuristics.
- Study 3 investigated the framing effect with positively and negatively presented information.
- Study 4 explored the endowment effect by comparing owned versus newly found items.
Main Results:
- ChatGPT demonstrated the anchoring effect, being influenced by random numerical anchors.
- ChatGPT exhibited representativeness and availability heuristics, overestimating joint probabilities.
- ChatGPT showed a framing effect, valuing items differently based on presentation.
- ChatGPT displayed an endowment effect, valuing owned items more than identical new items.
- Human participants exhibited similar heuristic and context-sensitive responses across all studies.
Conclusions:
- LLMs like ChatGPT can exhibit cognitive biases and context-sensitive responses, mirroring human behavior.
- The findings suggest that the structure and content of language may play a significant role in generating these effects.
- This research opens avenues for understanding the interplay between language, cognition, and AI behavior.
More Related Videos
06:48Lexical Decision Task for Studying Written Word Recognition in Adults with and without Dementia or Mild Cognitive Impairment
Published on: June 25, 2019
06:09P300-Based Brain-Computer Interface Speller Performance Estimation with Classifier-Based Latency Estimation
Published on: September 8, 2023
Related Concept Videos
Decision Making: P-value Method
First, a specific claim about the population parameter is proposed. The claim is based on the research question and is stated in a simple form. Further, an opposing statement to the claim is also stated. These statements can act as null and alternative hypotheses: a null hypothesis would be a neutral statement while the alternative hypothesis can...
Language and Cognition
Decision Making: Traditional Method
First, a specific claim about the population parameter is decided based on the research question and is stated in a simple form. Further, an opposing statement to this claim is also stated. These statements can act as null and alternative hypotheses, out of which a null hypothesis would be a...
Heuristics
People often rely on heuristics when faced with an overload of information, limited time, low importance of the decision, limited information, or when a heuristic readily comes to mind. For...
Decision Making
Automatic decision-making is fast, intuitive, and relies on gut feelings...
The Availability Heuristic