Generative Artificial Intelligence (GenAI), notably Large Language Models such as ChatGPT, has precipitated a fundamental transformation within higher education, redefining conventional pedagogical frameworks, knowledge acquisition methodologies, and assessment paradigms. This study comprehensively investigates the integration of GenAI in academic environments, focusing on its dual capacity to enhance personalized learning and its potential to undermine academic integrity. Employing a mixed-methods research design across multiple higher education institutions in Pakistan, the research analyzes the perceptions, usage patterns, and ethical concerns of both students and faculty members. The findings illuminate a complex dichotomy: while GenAI significantly augments research efficiency, writing proficiency, and personalized feedback, it concurrently introduces profound challenges regarding cognitive dependency, plagiarism, and the erosion of critical thinking skills. The study argues that prohibition is an untenable strategy; rather, institutions must proactively adapt curricula and assessment designs to foster ethical AI literacy. By synthesizing empirical data with contemporary theoretical frameworks, this paper provides actionable insights for educators and policymakers striving to balance technological innovation with rigorous academic standards in the rapidly evolving digital age.