James L. Waters Symposium: Accelerating Innovation in Analytical Chemistry and Measurement Science with Generative AI

To help analytical chemists become tomorrow's leaders, we must first examine the foundations of analytical sciences. Beyond chemical expertise, modern analytical chemistry rests on the pillars of applied physics and scientific computation. High-quality generative artificial intelligence (AI) can now serve as a tool for human innovators to discover, design, and develop more effective solutions to problems encountered in chemical analysis. Generative AI can also help reduce the perceived complexity of chemometric algorithms into manageable concepts that previously required months to master in the pre-AI era. This presentation proposes that generative AI can serve as a knowledgeable, Ph.D. level "colleague," especially for analysis problems that require advanced signal processing and chemometrics. These challenges include seeking scientific advice on analytical problems; using AI to deepen historical literature searches; developing feature-detection methods; performing information-theoretic denoising; enhancing resolution through regularized deconvolution, which can increase resolution; and determining instrument transfer functions in spectroscopy and separation sciences. Examples will show how generative AI can help distill the key ideas from complex code, rendering it into a readily understandable explanation for analytical chemists. The second part will focus on unconventional uses of ChatGPT (Pro), ethical issues, and its role as an assistant for chemical analysis tasks, ranging from diagnosing issues in publications to reading non-English analytical literature that is otherwise neglected due to language barriers. Together, these advances show that, when used critically and with appropriate benchmark checks, generative AI can accelerate innovation and self-education in analytical chemistry. Unverified use of generative AI can yield deceptive results that appear like the correct answers from Clever Hans, the horse once thought to do complex arithmetic a century ago.