If AI is sentient, then so is "Age of Empires II" (www.404media.co)

In a provocative new paper, Microsoft AI researcher Adrian de Wynter explores the absurdity of anthropomorphizing large language models (LLMs) by suggesting that if LLMs could be conscious, then so could the video game Age of Empires II. He created a simplistic neural network within the game using digital goats to illustrate his point, emphasizing that many researchers overlook the fundamental differences between LLMs and human-like consciousness. His work questions the widespread assumption that LLMs possess human attributes, which can skew experimental design and interpretation of results. De Wynter's findings call for a reevaluation of how the AI/ML community perceives LLMs. By debunking the notion that anthropomorphizing facilitates understanding, he urges scientists and executives to recognize that these models, while exhibiting complex behaviors, do not necessarily have consciousness. He highlights the marketing implications of this misconception, suggesting that emotional attachments to LLMs could drive product sales but ultimately cloud objective understanding. His research stands as a cautionary tale against conflating advanced algorithms with human-like qualities, fostering a more accurate discourse on the capabilities and limitations of artificial intelligence.