This paper explores the prevalence of biases in AI-driven educational technologies, examining their sources, manifestations, and impacts on diverse student populations. It proposes a framework for identifying and mitigating biases, emphasizing the need for transparent and inclusive educational technologies. The global market value of AI in education is projected to increase by over 45% CAGR(Compound Annual Growth Rate) between 2022 and 2030- highlighting the vast potential of this growing industry. As a responsible society, we must ensure the ethical reproduction of Artificial Intelligence in a sensitive domain like education, where the potential is boundless however the risks are equally potent. This paper aims to set the foundation of AI as an assistive technology and not as a replacement for traditional teaching. The research highlights the use of Deep Learning and Natural Language Processing in Curriculum Design or Pedagogy planning, focusing on critical thinking and not only static learning. While other recent works have merely reviewed one use case, we have categorized EdTech tools into three learning pathways: Student-Supporting, Teacher-Supporting, and System-Supporting. By theoretical and empirical analysis, we systematically break down two technologies Assistive Grading Technology and Chatbot Personalized learning, and consequently explore every impact on the stakeholders: students, teachers, and the educational community. Furthermore, we emphasize the need for a safe regulatory framework, much like the one currently existing, to ensure there is no misuse of data collected. The laws surrounding the fragility of the information sourced and the usage of data mining for further replication should be strict, with appropriate legal consequences for breaching the same.