Chemistry.AI - An AI Based Assistant for Chemistry Education
Timm Wilke, Carl von Ossietzky Universität Oldenburg, Institute for Chemistry, Chemistry Education Department, 26129 Oldenburg, Germany (Germany)
Abstract
As generative artificial intelligence (AI) becomes increasingly integral to educational reform, its capacity to revolutionize teaching cultures is becoming more apparent. In the realm of chemistry education, we develop an AI-based assistant to harness the potential of AI, particularly leveraging fine-tuned models based on OpenAIs large language model (LLM) / ChatGPT. This assistant is designed to address areas traditionally challenging for human educators, such as personalizing education, differentiating instruction, and automating grading and feedback processes. The final tool will focus on utilizing the strengths of AI to complement and enhance human capabilities in the educational sector. In our presentation, we provide an initial overview of our development process, highlighting how fine-tuned LLMs can significantly improve the quality of AI assistance in education. By exploring the implementation of this AI assistant in chemistry education, we aim to showcase its potential to facilitate a more tailored and efficient learning experience, thereby setting a new benchmark for the integration of technology in educational contexts.
Keywords: Artificial intelligence, teaching assistant