This paper presents a critical examination of ChatGPT’s role in education during the formative period from 2022 to 2025. It selectively analyzes key breakthroughs and enduring challenges associated with the integration of Generative Artificial Intelligence (Generative AI) into educational settings. Drawing from recent developments in pedagogy, educational technology, and institutional practice, the study identifies five major areas of innovation: AI-assisted tutoring, instructional content generation, feedback automation, support for creative inquiry, and just-in-time learning facilitation. These pedagogical advances are contextualized within the broader technological embedding of ChatGPT into learning management systems, productivity tools, and purpose-built educational applications. In parallel, the paper investigates critical risks, including threats to academic integrity, the erosion of critical thinking, epistemic dependence on AI-generated content, privacy concerns, and equity gaps in access to high-quality AI tools. The analysis demonstrates that while ChatGPT offers substantial benefits when thoughtfully integrated into pedagogical workflows, it also introduces systemic tensions that challenge traditional educational values and assessment practices. The paper concludes by arguing that future adoption of generative AI in education must be accompanied by institutional policies, AI literacy initiatives, and ethical frameworks that preserve the integrity of human-centered learning.