(261b) Sustainable Assessment of Emerging Technologies By Large-Language-Models-Based Genai Technology

The next generation of technological competition in manufacturing industries will be dictated by inventions. Although emerging technologies can become an engine of change and progress, the net profit brought to the environment and society could be questionable, if sustainability principles are not fully incorporated into technology development phases. This renders a need for comprehensive sustainability assessment of technologies in their different development stages. In this endeavor, a key step is to identify the technologies at different technology readiness levels, ranging from the basic research level to the ready-for-adoption level. In the metal and polymeric coating and surface finishing sectors, there exists numerous technical information openly accessible. Identification and classification of technologies by conventional search methods are time consuming, inefficient, and susceptible to errors.

The realm of Generative AI (GenAI) is considered on the cusp of a new transformative leap, transitioning from Large Language Models (LLMs), which are multi-billion parameter transformer neural networks that are trained on enormous collections of documents without supervision or labels. LLMs can perform multiple tasks like classifying natural language, translating text, and document search. Perhaps the most remarkable task of LLMs is to complete an input string of text. Via this mechanism, LLMs can write unit tests, document function, write code from a doc string, answer questions, etc. The recent big leap in AI-based assistant chatbots, GPT-4, has created new opportunities to automate engineering tasks and reduce the workload on human experts. This work focuses on harnessing GPT-4 for the identification of emerging technologies for sustainability improvement of surface finishing. Our investigation delves into the capabilities of this advanced model in executing tasks related to technology identification and classification. We will share our experience in generating chain-of-thought and maieutic prompts, exploring the benefits, pitfalls, and challenges of using it. The sustainability assessment and readiness levels of a resulting set of technologies will be also presented to show when and how these technologies could be used in the manufacturing sector.