What does this research mean for the field?

Integrating a large language model as a co-expert within a Delphi framework (AI-Delphi) effectively evaluates new analytical methodologies, demonstrating that Stop-Flow MicroScale Thermophoresis offers strong environmental benefits but is constrained by analytical and operational limitations. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.ESTABLISHES_NEW_DIRECTION.

December 8, 2025Open Access

Can artificial intelligence assist experts in evaluating new analytical methodologies? Stop-flow MicroScale Thermophoresis assessed using the RGB model and AI-Delphi framework

PNPaweł Mateusz NowakJagiellonian University MWMichał WoźniakiewiczJagiellonian University AWAneta WoźniakiewiczJagiellonian University

Key Points

Key points are not available for this paper at this time.

Abstract

• SF-MST was evaluated using the AI-Delphi framework. • Independent experts iteratively assessed SF-MST via RGBₑx-ante model. • ChatGPT synthesized experts input and provided its own opinion. • SF-MST shows increased greenness, but worse analytical and practical potential. • Perspective of using AI in method evaluation is discussed. Stop-Flow MicroScale Thermophoresis (SF-MST) enables thermophoretic analysis of highly miniaturized samples introduced into ultrathin capillaries by pressure or electrophoresis. Although successfully applied by our group in several research areas, the broader analytical potential and overall added value of SF-MST remain uncertain and require structured evaluation. To address this, the RGBₑx-ante model based on the White Analytical Chemistry (WAC) concept was applied to prospectively assess this technique in terms of analytical performance (red), environmental friendliness and safety (green), and practicality (blue), which together determine its overall “whiteness”. Seven independent experts participated to ensure objectivity, and moreover, generative artificial intelligence (ChatGPT) was incorporated as an element of the iterative consensus process (AI-Delphi). The final whiteness score obtained from experts, after considering both peer and AI assessments, was 90. 5, indicating a moderate potential of SF-MST. The technique offers clear environmental and conceptual benefits, but its analytical and operational limitations still restrict broader implementation. This work presents the first integration of the RGB model with a Delphi-type expert framework and the first use of a trained large language model as a co-expert in the evaluation of whiteness. The study not only clarifies the strengths and weaknesses of SF-MST but also suggests how AI-supported assessment can enrich the methodological landscape of Green and White Analytical Chemistry.

AIに質問

Bookmark

View Full Paper