Predicting Biased Human Decision-Making with Large Language Models in Conversational Settings
📰 ArXiv cs.AI
Learn how large language models can predict biased human decision-making in conversational settings and capture cognitive biases under cognitive load
Action Steps
- Run a pre-registered study to collect data on human decision-making in conversational settings
- Use large language models to predict biased decision-making and capture cognitive biases
- Apply cognitive load to participants and analyze how it affects decision-making
- Configure a chatbot to administer decision-making tasks with varying complexity
- Test the predictions of the large language models against actual human decision-making data
Who Needs to Know This
Data scientists and AI researchers can benefit from this study to improve their understanding of human decision-making and develop more accurate predictive models. Product managers can also apply these insights to design more effective conversational interfaces
Key Insight
💡 Large language models can capture not only human cognitive biases but also how those effects change under cognitive load
Share This
🤖 Large language models can predict biased human decision-making in conversations! 📊
Key Takeaways
Learn how large language models can predict biased human decision-making in conversational settings and capture cognitive biases under cognitive load
Full Article
Title: Predicting Biased Human Decision-Making with Large Language Models in Conversational Settings
Abstract:
arXiv:2601.11049v2 Announce Type: replace-cross Abstract: We examine whether large language models (LLMs) can predict biased decision-making in conversational settings, and whether their predictions capture not only human cognitive biases but also how those effects change under cognitive load. In a pre-registered study (N = 1,648), participants completed six classic decision-making tasks via a chatbot with dialogues of varying complexity. Participants exhibited two well-documented cognitive bias
Abstract:
arXiv:2601.11049v2 Announce Type: replace-cross Abstract: We examine whether large language models (LLMs) can predict biased decision-making in conversational settings, and whether their predictions capture not only human cognitive biases but also how those effects change under cognitive load. In a pre-registered study (N = 1,648), participants completed six classic decision-making tasks via a chatbot with dialogues of varying complexity. Participants exhibited two well-documented cognitive bias
DeepCamp AI