Responsible AI in Energy System Analysis: Methodology, Transparency, and Standards – VKI-Energy

In the VKI-Energie project, the Reiner Lemoine Institute is investigating how artificial intelligence methods can usefully complement energy system analysis and how their application can be made transparent, traceable, and scientifically sound. The project focuses on two key areas: a structured overview of AI application fields in energy system analysis and an evaluation framework for the responsible use of AI.

Systematically classifying AI applications

The researchers are building on the results from the preliminary project KIRLI. They are categorizing AI methods with a clear connection to the research fields of energy system analysis. They are reviewing the literature and existing preliminary work, clustering application areas such as forecasting, optimization, data processing, and knowledge representation, and linking them to potential AI methods. The goal is to identify where AI can usefully complement or replace existing methods of energy system analysis.

From an overview to building internal expertise

The result is a summary table specific to energy system analysis. It is intended to provide guidance for future research and consulting projects and to support the development of expertise at the institute as well as in other research organizations. The researchers are developing materials for informational formats and public outreach.

Assessing Transparency, Quality, and Accountability

In addition, the researchers in the project are developing a transferable evaluation framework for AI applications in energy system analysis. The focus is on criteria for quality, robustness, fairness, explainability, reproducibility, and documentation of the applications. The researchers are reviewing existing standards and frameworks, evaluating them, and translating them into a concept for technical metrics, documentation standards, and transparency criteria.

Guidelines for Future Projects

The results of this evaluation framework will be incorporated into guidelines for the responsible use of AI in energy systems research. These guidelines are intended to enhance the credibility of AI-supported results, provide guidance for follow-up projects, and contribute to the standardization of transparent AI applications in energy research. RLI will publish the evaluation framework and present it at workshops, conferences, and other communication events.

Project period: April 2026 – December 2026

• Systematic classification of relevant AI methods for applications in energy research, particularly ontologies and open data structures
• Development of a concise overview of the potential applications, added value, and limitations of AI methods
• Research and derivation of criteria for transparent, robust, and traceable AI use
• Development of material for a transferable evaluation framework for responsible AI in energy systems research
• Preparation of the results for standardized knowledge formats and external communication