Enhancing accessibility for blind and visually impaired (VI) students at universities is essential for promoting equitable participation in academic and social life. This study examines the efficacy and limitations of navigation systems powered by artificial intelligence (AI) within academic environments, with a particular emphasis on improving accessibility for students who are blind or visually impaired. The results demonstrate that although AI-driven systems exhibit significant accuracy in localization and obstacle avoidance, their application in varied, real-world university environments is still insufficiently investigated, especially in developing areas like Bangladesh. Significant challenges, such as adapting to changing environments and addressing cost factors, are highlighted, along with prospective research avenues that include broader user studies and the development of open-source solutions. This study emphasizes the significant impact of artificial intelligence on improving accessibility within educational institutions and calls for additional exploration to address current research deficiencies.