Do Large Language Models Shop Like People?

In a recent study published in Food Quality and Preference, we investigated whether large language models can reproduce human food choices in structured market-like settings. The paper, titled “Do large language models shop like people? Comparing GPT and human food choices using discrete choice experiments”, compares GPT-generated choices with human choices across three food-related discrete choice experiments: 3D-printed chocolate, tomato sauce from social farming, and fresh tomatoes with environmental quality claims. Our results show that GPT can reproduce the broad qualitative structure of human preferences, often identifying the relevant product attributes and the expected direction of their effects. However, the […]