Not Minds, but Signs: Reframing LLMs Through Semiotics [pdf] (arxiv.org)

A recent paper titled "Not Minds, but Signs: Reframing LLMs Through Semiotics" argues for a fundamental shift in how we understand Large Language Models (LLMs). Instead of viewing them as cognitive entities that simulate human understanding, the authors advocate for a semiotic perspective. This perspective emphasizes that LLMs primarily operate by recombining and recontextualizing linguistic elements based on probabilistic relationships, effectively positioning them as agents within cultural processes that generate texts for interpretive engagement rather than cognitive comprehension. This reframing is significant for the AI/ML community as it challenges anthropomorphic notions surrounding LLMs and underscores their role as tools that facilitate creativity and critical inquiry. By analyzing how LLMs function as semiotic agents, the paper opens new avenues for applications in literature, philosophy, and education, highlighting their capabilities to influence meaning-making within socially embedded contexts. The semiotic approach encourages a more ethically aware interaction with these technologies, reshaping our understanding of language, interpretation, and the impact of artificial systems on knowledge production.