On 22 October 2026, Prof. Bernard Slippers led a hands-on workshop at TU Braunschweig in Germany on the use of AI in Science, with a focus on plant-associated microbiomes. The half-day workshop brought together researchers from TU Braunschweig, including the group of Prof. André Fleißner at the Institute of Genetics, and colleagues from the University of Hamburg, the Julius Kühn-Institut, the Leibniz Institute DSMZ, and FABI (not all represented on the photo). It formed part of Bernard's Alexander von Humboldt Foundation Research Award stay at TU Braunschweig.
The workshop opened with a discussion of AI in science, including security and ethics. Participants then set up and worked with AI research tools on their own laptops, using folders of papers and documents from their own projects. Bernard shared what he has learned over the past year from using these tools to analyse plant-associated microbiomes, amongst other topics. A recurring theme was the need for rigour, structure and systematic verification in AI-assisted analysis.
The second part focused on the microbiome analysis pipeline Bernard has developed, how to use it, and how participants could build their own. He also introduced the plant microbiome meta-analysis resource he is doing that brings together public bacterial and fungal datasets while keeping each study's own design intact, and invited participants to contribute, for example by checking the records of studies they know well. The discussion also returned to what we should train researchers for: knowing the organism, designing the measurement that would settle a question, and critically interpreting the results.
The workshop builds on similar training Bernard has presented [this year for AATF, Stellenbosch University and FABI](https://fabinet.up.ac.za/news/mastering-ai-research-assistants-reflections-from-a-sabbatical). Participants received a general scaffold for building such research assistants, and Bernard hopes to build a cohort of interested scientists to develop this work further.