Pangram’s CEO, Max Spero, claims that language models can be fooled by making the same arguments, as the tool detects suspicious phrases but struggles to understand the underlying patterns. Spero argues that while LLMs excel at grammar and logic, their uniformity limits their ability to produce diverse arguments. This highlights the need for better arguments to outsmart language models. Why it matters: The limitations of language models underscore the importance of developing more nuanced and human-like AI capabilities.