The presentation addresses the epistemological juncture created by the application of Artificial Intelligence (AI) to animal communication studies. It aims not to reject these technologies, but to critically reframe them within the robust theoretical framework of biosemiotics, arguing that the prevailing metaphor of “translation” or “decoding” is epistemologically misleading. The methodology involves a critical analysis of the computational paradigm, examining projects like CETI and ESP to reveal their implicit assumptions: a view of communication as a formal, decontextualized code modeled on anthropocentric linguistics. Subsequently, the paper employs key biosemiotic concepts, such as Umwelt, embodiment, and the triadic model of the sign, to demonstrate that living semiosis is an embodied, situated, and relational process irreducible to purely syntactic analysis. The conclusion posits that biosemiotics is not an accessory but a necessary condition for a non-reductionist understanding. It proposes reconfiguring AI not as a translator, but as a powerful heuristic tool for revealing patterns that must be interpreted through a deep hermeneutics of animal worlds, fostering a science of listening over the hubris of decipherment.

– Nicola Zengiaro is a biosemiotician and a postdoc fellow in his Marie Curie Global project “VITAE” at Charles Univeristy, in Prague, in collaboration with the Universidad EAN in Bogotá.