Comparing large language model artificial intelligence tools in aid of electrical engineering
Zoltán Tafferner, Attila Géczy · 2024
Conversational AI (artificial intelligence) systems have broken into the digital technology sphere at the end of 2022. Many scientists, engineers, and authors have begun using them despite a lack of evidence for their applicability. Research has been conducted in several fields, including electrical engineering, on the applicability of generative AI in many fields, such as chemical engineering, programming, medicine, and even electrical engineering. Most, although not all, of the existing research focuses on applying specific AI models. This case study compares several generative AI models (ChatGPT, Bard, Bing AI, Copilot) through tasks in electrical engineering design. We qualitatively evaluated the AI responses to the questions. We have found that although AI tools are applicable and perform similarly well in general questions, their use is highly questionable in more specific cases, for example, in circuit design. Lastly, AIs perform equally unacceptably in tasks that require specific data, such as literature reviews. We conclude that the bots should be used sparingly with strict supervision and professional knowledge.