Year: 2026
Type: review
Abstract: Generative artificial intelligence can design biological sequences, from proteins to whole genomes, opening opportunities for medicine and biotechnology, but the same capability can be misused. The Wübben Sandpit conference brought together experts in biology, AI, biosecurity, ethics, medicine and policy to ask how the field can capture the benefits while managing the risks. Here we set out a defence-in-depth strategy that builds safeguards into every stage, from the data used to train models, to the models themselves, to the synthesis of physical DNA. We argue that responsible governance is not a brake on progress but a precondition for it. Generative artificial intelligence can design biological sequences, from proteins to whole genomes, however this capability can be misused. Here the attendees of the Wübben Sandpit conference outline a defence-in-depth strategy that builds safeguards into every stage of sequence design.
Source: Nature Communications
Authors: Maximilian Sprang, Hassan Ahmed Hassan, Yasha Ektefaie, Anne Schmieder, Alix Wasilenko, Moritz S. Hanke, Rosa E. Martín-Peña, Rachael Harkness, Jasper Götting, Alan Murphy, Johannes U. Mayer, Marco Schäfer, André Friedrich, Kerstin Lenhof, Sophia Rudorf, Gigi Gronvall, Jan Kellmann, Simon Anders, Berkan Aysan, Nicole Saverschek, Kathrin Schuldt, Marc Güell, Katharina Ladewig, Georges Hattab
Institutions: Johannes Gutenberg University Mainz, University Medical Center of the Johannes Gutenberg University Mainz, Broad Institute, Harvard University, Center for Scalable Data Analytics and Artificial Intelligence, Technische Universität Dresden, Leipzig University, University of Zurich, Johns Hopkins University, Leibniz University Hannover, CAIMed - Niedersächsisches Zentrum für KI und Kausale Methoden in der Medizin, University of Edinburgh, Cold Spring Harbor Laboratory, Robert Koch Institute, Universitätsmedizin Göttingen, Philipps University of Marburg, Loewe Center for Synthetic Microbiology, Heidelberg University, Stiftung der Deutschen Wirtschaft, Bernhard Nocht Institute for Tropical Medicine, Universitat Pompeu Fabra, Medizinische Universität Lausitz - Carl Thiem, Freie Universität Berlin
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $6,990
Award: G6122989955
OpenAlex ID: W7165888489
