Chatbot Artificial Intelligence (AI) to Supporting Differentiated Mathematics Learning

Artificial Intelligence chatbots have become famous nowadays because of their ease of use, which is even felt by educators and students. However, the AI chatbot algorithm sometimes does not match what is expected because it is not precise and does not match the prompt. For example, when we type the prompt "basic concepts of mathematical modeling for middle school class VII" there is very little explanation of the basic concepts. The search results provided by AI chatbots are too complex for students who want to learn simple basics. Based on this, this research was carried out with the aim of developing an AI chatbot to support differentiation learning. This type of research is ADDIE model development research which is carried out in five stages, namely analysis, design, development, implementation and evaluation. The results of the research at the analysis stage are that students need learning media that is appropriate to their level of understanding of mathematical modeling material as basic material before studying other material. The results of this analysis are the basis for designing an AI chatbot and bot scenarios to be developed and given to experts, namely educational technologists and mathematics education lecturers. After that, small class trials and large class trials at the implementation stage to get practicality and effectiveness test results in using the AI chatbot. The final stage is evaluation as a process of reflection on the results of AI chatbot development. The research results show that the AI chatbot media developed is valid, practical and effective. Valid criteria with a percentage of 89.7%, practical criteria of 85% and effective criteria of 77.67% in small group trials. Keywords: chatbot AI, differentiated, mathematics learning.