Humans & LLMs rate deliberation as superior to intuition on reasoning tasks (www.nature.com)

A recent study involving 13 experiments with over 2,500 participants revealed a strong preference among humans for deliberative reasoning over intuitive decision-making, particularly in complex reasoning tasks. Participants consistently rated individuals who engaged in thoughtful deliberation as smarter and more trustworthy, even when the accuracy of their decisions was equal to that of their intuitive counterparts. This pattern of preference was mirrored in responses generated by large language models (LLMs) like ChatGPT 3.5 and 4, suggesting that these models reflect common human beliefs about reasoning styles. Even under conditions that typically demand quick decision-making, like time pressure, the valuation of deliberative thought prevailed. This research holds significant implications for both the AI and ML communities, highlighting the need for LLMs to incorporate more deliberative reasoning processes to enhance their credibility and public acceptance. As AI systems increasingly recommend decisions in areas such as healthcare and finance, aligning their outputs with human preferences for deliberation could foster greater trust in AI-generated advice. With these insights, developers can refine prompts and computational techniques to model deliberative reasoning more effectively, ultimately bridging the gap between human intuition and AI capabilities.