Surveys and Artificial Intelligence: A New Challenge for Non-market Economic Valuation

The widespread adoption of Artificial Intelligence (AI) presents substantial challenges within the field of economics, prompting a discussion on the concepts of homo economicus and machina economicus . The use of AI-generated Silicon samples adds a new dimension to non-market economic valuation. Machina economicus may be capable of articulating preferences and elucidating decision-making processes, evoking human behavior. Discrete choice surveys are nowadays mainly conducted using internet panels, and several concerns have been raised regarding the quality of responses and representativeness of these samples. The main aim of this paper is to analyze preferences from a Silicon sample, and compare the results with a study based on responses of people from an internet sample. The results show that estimates of willingness to pay in the Silicon sample range between 5.34 and 15.70 times those of the internet sample. Silicon samples can be helpful in the initial stages of designing a discrete choice experiment, but they do not replace real people.