Integration of Generative AI for Intelligent Diagnostic of Vehicles

The latest update on Generative AI, especially the OpenAI platform, confirms they are no longer single-domain proficiency. It already has been adopted by a variety of applications, emphasizing speech, audio, vision, and multimodal interactions. GPT family of large language models (LLM) is evolving where GPT4o is a few times faster, way more intelligent, and at the same time much cheaper as a service compared to the previous version. Additional services such as “AI Assistant” make it possible to be utilized with fewer resources available in the embedded systems. The advancement of semiconductor technology and microelectronics in the form of Wireless connected SoC in small sizes opens many fields for various IoT (Internet of Things) applications. The proposed solution in this paper will present a system that combines those two technologies in the form of a standalone embedded connected device, that can perform autonomously diagnostic of vehicle by identifying and prompting the owner or driver for troubles in the vehicle and efficiently suggesting for corrective actions. The results show how comprehensive troubleshooting of logged errors is with a standalone embedded system by the support only of GPT AI Assistant.