Tech giants in Silicon Valley, including OpenAI and Anthropic, have been advocating the use of AI to tackle complex mathematical problems at the highest levels. This move, however, has sparked apprehension among mathematicians, notably Li Xinyi, a researcher at Peking University. Li contends that AI-driven solutions predominantly rely on amalgamating pre-existing human knowledge bases, rather than generating novel, human-comprehensible theories. He argues that the extensive promotion of such AI capabilities misleads the public and those responsible for funding decisions, potentially dissuading young talents from pursuing mathematics. Furthermore, he raises concerns about the risk of AI appropriating unpublished human research findings. Li underscores the importance of preserving the community of human mathematicians, safeguarding the传承性 (heritable and transmissible) system of human mathematical knowledge, and ensuring that academic accolades are not exclusively bestowed upon individuals who merely craft AI prompts. At present, AI still faces significant challenges in conquering mathematical domains that heavily rely on human intuition.