Predicting Biased Human Decision-Making with Large Language Models in Conversational Settings
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 biases: the Framing Effect and the Status Quo Bias. Increased dialogue complexity resulted in participan
The rapid advancement and widespread deployment of large language models are creating an urgent need to understand their capabilities and limitations in emulating and predicting human behavior.
Understanding how LLMs can predict and potentially exploit cognitive biases in conversational settings has profound implications for human-AI interaction, ethical AI development, and digital persuasion.
This research demonstrates LLMs' potential to not only replicate cognitive biases but also to model how these biases are affected by external factors like cognitive load, suggesting a new dimension in AI's understanding of human decision-making.
- · AI ethicists
- · UX designers
- · Behavioral scientists
- · LLM developers focused on human-centric AI
- · Consumers susceptible to digital persuasion
- · Developers of un-auditable AI systems
- · Organizations relying on unchecked AI 'agents'
LLMs can be trained or refined to better anticipate and possibly influence human decision-making in interactive environments.
This capability could lead to more sophisticated AI agents designed to guide users, for better or worse, through complex decisions by leveraging predicted biases.
A potential future where LLM-powered interfaces are so adept at predicting and exploiting cognitive frailties that human autonomy in digital spaces is significantly eroded.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.AI