压力下的星球人工智能

Title:AI for a Planet Under Pressure Authors:Victor Galaz,1,2*Maria Schewenius,1,2,6*Jonathan F. Donges,1,3,5Ingo Fetzer,1,2,7Erik Zhivkoplias,1Wolfram Bar-fuss,8,9,10Louis Delannoy,1,11,12Lan Wang-Erlandsson,1,3Maximilian Gelbrecht,3,13Jobst Heitzig,3Jonas Hentati-Sundberg,14Christopher Kennedy,15Nielja Knecht,1Romi Lotcheris,1Miguel Mahecha,16Andrew Merrie,1,24David Montero,16TimonMcPhearson,1,15Ahmed Mustafa,15Magnus Nyström,1Drew Purves,4Juan C. Rocha,1Masahiro Ryo,17,18Claudia van derSalm,4Samuel T. Segun,19Anna B. Stephenson,20Elizabeth Tellman,21Felipe Tobar,22Alice Vadrot23 *Editors Affiliations: 1.Stockholm Resilience Centre, Stockholm University, Sweden2.Beijer Institute of Ecological Economics, Royal Swedish Academy of Sciences, Sweden3.Potsdam Institute for Climate Impact Research (PIK), Member of the Leibniz Association, Germany4.Google DeepMind, United Kingdom5.Max Planck Institute of Geoanthropology, Germany6.Faculty of Engineering and Sustainable Development, University of Gävle, Sweden7.Bolin Centre for Climate Research, Stockholm, Sweden8.Transdisciplinary Research Area Sustainable Futures, University of Bonn, Germany9.Center for Development Research, University of Bonn, Germany10.Institute for Food and Resource Economics, University of Bonn, Germany11.Global Economic Dynamics and the Biosphere, Royal Swedish Academy of Sciences, Sweden12.Swedish Centre for Impacts of Climate Extremes (climes), Uppsala University, Sweden13.School of Engineering and Design, Technical University of Munich, Germany14.Swedish University of Agricultural Sciences, Sweden15.Urban Systems Lab, New York University, USA16.Institute for Earth System Science and Remote Sensing, Leipzig University, Germany17.Brandenburg University of Technology Cottbus-Senftenberg, Germany18.Leibniz Centre for Agricultural Landscape Research (ZALF), Germany19.Global Center on AI Governance, Canada20.High Meadows Environmental Institute, Princeton University, USA21.Nelson Institute for Environmental Studies, University of Wisconsin-Madison, USA22.Department of Mathematics, Imperial College London, United Kingdom23.Department of Political Sciences, University of Vienna, Austria24.Planethon, Sweden Suggested citation for the full report:Galaz, V. and M. Schewenius (eds, 2025).AI for a Planet Under Pressure.Stockholm Resilience Centre, Potsdam Institute for Climate Impact Research. Stockholm. Report. Online: http://arxiv.org/abs/2510.24373. Suggested citation for individual chapters in report (e.g.):Wang-Erlandsson, L., N. Knecht, R. Lotcheris, I. Fetzer. “Securing Freshwater for All” in Galaz, V. and M. Schewenius (eds, 2025).AI for a Planet Under Pressure.Stockholm Resilience Centre, Potsdam Institute for Climate Impact Research.Stockholm. Report. ISBN:978-91-89107-61-8e-book ISBN:978-91-89107-62-5 Graphics and layout:Jerker Lokrantz/Azote November 2025 The production of this report was supported by additional funding from the Beijer Institute of Ecological Economics (Royal Swedish Academy ofSciences), and Google.org through ClimateIQ at the Urban Systems Lab at New York University (USA). Executive Summary Artificial intelligence (AI) is already driving scien-tific breakthroughs in a variety of research fields,ranging from the life sciences to mathematics.This raises a critical question: can AI be appliedboth responsibly and effectively to addresscomplex and interconnected sustainabilitychallenges? These challenges include climatechange, biodiversity loss, ocean acidification,and other transformations of our living planet.Each of these poses serious risks to societalstability, human health and well-being, and theability of present and future generations to thrivewithin planetary boundaries. Here, science playsa fundamental role in two ways.First, by helpingaccelerate changes and innovations that moveus closer to a just and safe future for all.Second,by making sure that AI is developed and used inways that neither exacerbates inequalities,nor increases planetary pressures. ic publications, and expert deep-dives into eachspecific issue area. In conclusion, we show that:(1) AI offers vast potential to accelerate pro-gress across the sustainability sciences. (2) AIcan sharpen our decision-making and clarifycomplex environmental challenges for research-ers and the public alike. (3) However, realizingthis promise requires careful navigation of therisks, including AI’s own environmental footprint,inherent biases, and the challenge of unequalaccess. (4) Despite these hurdles, responsibleand ethical applications of AI in sustainabilityresearch are not just a possibility—they are anurgent necessity. (5) Pioneering these uses canunlock the breakthroughs we need to build amore sustainable future. Advancing from the potential of AI to its respon-sible application in research that benefits bothpeople and the planet will require a balancebetween urgency and innovation, and a com-mitment to ethical and inclusive practices. Ourconcluding