Working Paper
Recent studies show that large language models (LLMs) can augment the generation and evaluation of ideas, but their impact on problem formulation remains underexplored. Through a randomized controlled trial with 305 MBA students, the authors investigate how LLM assistance at different stages of the decision process affects strategic outcomes. Consistent with prior findings, the authors find that LLM assistance increases the number of alternatives generated. However, a surprising pattern also emerged: using LLMs in both problem formulation and ideation decreases strategic focus, an effect not observed when LLMs assist only in ideation. Using an abductive approach, the authors propose that when introduced during problem formulation, LLMs appear to shape how individuals mentally construct the strategic problem and perceive LLM’s role, creating cognitive anchors that bind subsequent search.
This study advances the understanding of human-AI collaboration in strategic contexts, highlighting the importance of when and how LLM is integrated in decision-making.
