Given such diverse fields of application, one might ask: What does AI mean for the disposal of nuclear waste? The Federal Office for the Safety of Nuclear Waste Management (BASE) is investigating this question.

What artificial intelligence actually means for nuclear waste disposal, however, is not an easy question to answer. The term itself is not clearly defined. Some definitions state that AI emulates human thought processes, others say this doesn’t matter. Some definitions claim that AI surpasses human performance. Others settle for AI being equivalent to human performance. Moreover, the term AI encompasses a variety of methods.

Specific use cases in nuclear waste management where the application of artificial intelligence is conceivable must, therefore, be identified first. One example of this is the AI-supported evaluation of data on geological conditions for the question of final disposal of radioactive waste.

Once such use cases have been identified, the use of AI raises elementary questions, especially and even more so in the risk-laden area of nuclear waste disposal. BASE is particularly interested in the following issues:

1. What questions arise for the regulation of AI when it is used by operating companies in nuclear waste management? What quality and transparency requirements are needed, for example, when an applicant in a supervision or licensing procedure uses AI?
   
   
2. What applications might BASE itself make use of within the framework of its specialist tasks? How could AI be used, for example, in the context of public participation? How could AI support BASE in supervisory reviews of extensive documents and data volumes?

BASE publication on artificial intelligence

BASE organised a transdisciplinary panel entitled "Quo Vadis Artificial Intelligence in Nuclear Waste Management" on 1 December 2022. Experts discussed and reported on the status of AI developments in the field of nuclear waste management.

The event brought together scientific expertise and practical experience to confront theoretical potential with the conditions of practice. The speakers included experts from political science, geography, computer science and law. The practical side was illuminated by presentations from the Federal Environment Agency and the Federal Radioactive Waste Disposal Company, among others.
A BASE publication makes the debates that took place there accessible to the public. An introductory overview of the topic precedes the following speaker contributions:

• Judith Krohn provides an overview of the fields where research into AI methods for dealing with geodata has already been carried out.
• Manfred Krafczyk illustrates how AI methods support the development and application of computational models and simulations. He goes on the explain how such AI methods can potentially be of use in the field of nuclear waste disposal, too.
• Vinzenz Brendler elaborates on the added value of the 'digital twin' AI method for nuclear waste disposal.
• In her contribution, Julia Romberg explains how AI methods can support citizen participation processes.
• Stephanie Schiedermair and Johannes Weil reflect on legal challenges that arise from the use of AI in public administrations.
• Annika Froese and Ina Richter conclude the publication with a consideration from BASE's point of view. They summarise, once again, the potentials and limitations of AI methods in the field of nuclear waste disposal.

Other BASE activities on the topic

• The research project "Application of artificial intelligence (AI) for the site selection of deep geological repositories"(in German) evaluates the application of AI in geoscientific issues.
• BASE cooperates with the Application Laboratory for Artificial Intelligence and Big Data at the Federal Environment Agency (UBA), and is a member of the BMUV's AI Task Force.

If you have any queries or require further information, please contact: