Recently, OpenAI, the renowned Open Artificial Intelligence Research Center headquartered in the United States, unveiled a collection of over 700 mathematical manuscripts crafted by its proprietary models. The organization boldly proclaimed that these models have successfully tackled more than 300 previously insurmountable mathematical conundrums spanning diverse domains, including algebra and logic. This revelation has since captured the spotlight within academic circles. While some herald this as a harbinger of transformation in the mathematical landscape, a significant number of mathematicians remain unconvinced. They underscore the immense verification burden associated with this batch of manuscripts, noting that three have already been retracted post-publication owing to inaccuracies. Furthermore, certain proof techniques generated by AI deviate from conventional human reasoning patterns.