The world was greatly amazed when ChatGPT, a large language model (LLM), became available. ChatGPT impacted many areas, especially academic research. This paper studies how ChatGPT can be very useful but also has some limits, especially in research. While ChatGPT is very good at creating clear and relevant text, it has biases from its training data, so results must be checked carefully. Also, it has problems with non-written data. More research and development are needed to fully understand how ChatGPT can change research practices and bring new ideas.