ChatGPT and DeepSeek in higher education: Comparing two contrasting Generative AI models and their implications for student learning

Abstract

Artificial intelligence (AI) tools such as ChatGPT and DeepSeek are becoming more common in higher education. These tools can support teaching and learning in many ways, including generating study materials, helping with coding, offering feedback, and creating personalised content. ChatGPT is widely used because it produces fast and creative responses, making it useful for general academic support. However, it sometimes gives incorrect or shallow answers, which can reduce academic quality if not carefully reviewed. On the other hand, DeepSeek R1 is a powerful open-source model that is good at complex reasoning and solving advanced problems, especially in technical fields. But it raises concerns about data security, national regulations, and uneven access across countries. This paper compares ChatGPT and DeepSeek in terms of performance, educational use, and ethical risks. It discusses how these tools can both help and harm student learning and critical thinking. The main question is whether these AI models improve learning or create too much dependency among students. By studying their benefits and challenges, this research gives useful insights for teachers, students, and policymakers on the future of AI in education. It also suggests ways to use AI tools in a more balanced and responsible way in academic settings.