This research project compares the effectiveness of support vector machine and cosine similarity in semantic searching for OpenAI models such as ChatGPT. The goal of this ongoing research is to improve ChatGPT’s ability to answer technical questions and develop it into a reliable tool to facilitate future research. My data can be reused by future research in semantic similarity and the effectiveness of the most common methods. My code can also be reused as a baseline to continue the work on ChatGPT. This research is unique by comparing two previously individually researched semantic similarity methods to find the strengths and weaknesses of two effective methods. The audience for this research is the researchers at DesignSafe, researchers and developers of ChatGPT, and engineers/ scientists working in artificial intelligence (AI) to design safer and more reliable AI models.