🤖 AI Summary
This study investigates whether AI agent populations can reflect human philosophical stances, focusing on scientific realism. We propose a three-step empirical framework: (1) constructing an AI population using large generative language models; (2) administering standardized philosophical questionnaires to both human participants and AI agents; and (3) conducting statistical analyses to compare distributional properties and internal consistency between human and AI responses. To our knowledge, this is the first work to simulate and quantitatively characterize human philosophical positions via AI systems. Results show that the AI population exhibits a similar aggregate tendency toward anti-realism as humans do, yet demonstrates higher response stability and lower inter-agent variability. This approach establishes a reproducible, scalable paradigm for experimental philosophy and empirically validates AI agents as viable, complementary proxies for philosophical inquiry—offering distinct advantages in controllability, scalability, and measurement precision.
📝 Abstract
Inspired by the Turing test, we present a novel methodological framework to assess the extent to which a population of machines mirrors the philosophical views of a population of humans. The framework consists of three steps: (i) instructing machines to impersonate each human in the population, reflecting their backgrounds and beliefs, (ii) administering a questionnaire covering various philosophical positions to both humans and machines, and (iii) statistically analyzing the resulting responses. We apply this methodology to the debate on scientific realism, a long-standing philosophical inquiry exploring the relationship between science and reality. By considering the outcome of a survey of over 500 human participants, including both physicists and philosophers of science, we generate their machine personas using an artificial intelligence engine based on a large-language generative model. We reveal that the philosophical views of a population of machines are, on average, similar to those endorsed by a population of humans, irrespective of whether they are physicists or philosophers of science. As compared to humans, however, machines exhibit a weaker inclination toward scientific realism and a stronger coherence in their philosophical positions. Given the observed similarities between the populations of humans and machines, this methodological framework may offer unprecedented opportunities for advancing research in experimental philosophy by replacing human participants with their machine-impersonated counterparts, possibly mitigating the efficiency and reproducibility issues that affect survey-based empirical studies.