Leiden Declaration on Artificial Intelligence and Mathematics

Signatories

The primary signatory list now lives on the homepage below the declaration text.

Sign Declaration
4176 Signatories
  1. Jean Avan ORCID CNRS, Laboratoire de Physique Théorique et Modélisation, Cergy
  2. MN Prashanth Kumar Student at Indian Statistical Institute, Bangalore
  3. Laura Kanzler verified email CNRS - LJLL, Sorbonne Université
  4. Colin Jakob Rittberg ORCID Postdoctoral Researcher, Vrije Universiteit Brussel
  5. Willi More University of Klagenfurt
  6. Arkaraj Mukherjee Student, Indian Statistical Institute
  7. Amaury Hayat ORCID Professor, École des Ponts - Institut Polytechnique de Paris
  8. Jose Antonio Ribeiro Neto ORCID
  9. Rajula Srivastava verified email University of Wisconsin - Madison
  10. Robin Bulsink verified email University of Groningen
  11. Maike Meier ORCID Assistant Professor, University of Groningen
  12. James Matthew Jones ORCID Postdoc Research Assistant, Philipps-Universität Marburg
  13. Pascal Romon ORCID Maître de Conférences, Université Gustave Eiffel
  14. Josué Tonelli-Cueto ORCID CUNEF Universidad
    Comment

    Joining collective efforts, specially in regard to new disruptive technologies such as AI, is a must if we want to align these new technologies with our community-values. Although the declaration (and its signing) by itself is not enough, signing serves as a strong signal regarding the values of the mathematical community towards the rest of the world and it is a first step towards more radical actions.

  15. Franck Corset ORCID DATA, Laboratoire Jean Kuntzmann
  16. Peter Bank verified email Technische Universität Berlin
  17. Pawel Strzelecki ORCID University of Warsaw
  18. Nathan Clisby ORCID Swinburne University of Technology
  19. Elisabeth Gassiat ORCID Université Paris-Saclay
  20. Giorgio Mangioni ORCID PhD Student, Heriot-Watt University
    Comment

    We should also think of implementing tools to make sure only authorised bots (e.g. the WayBack machine) can scrape journals and ArXiv

  21. Mark Kamsma ORCID Marie Skłodowska-Curie Postdoctoral Fellow, Masaryk University
  22. Veronika Kuchta ORCID Assistant Professor, Florida Atlantic University
  23. Georgios Politopoulos Leiden University
  24. Richard Ren ORCID Department of Mathematics, University of Chicago
  25. Paulo Félix Lamas ORCID Universidade de Santiago de Compostela
  26. Flavio Salizzoni ORCID Max Planck Institute for Mathematics in the Sciences
  27. Pip Goodman ORCID Universitat de Barcelona
  28. Rodrigo Thomaz ORCID PhD Candidate, Universidade Federal de São Carlos/Université de Lille
  29. Emmanuel Maitre ORCID Grenoble INP - Ensimag, UGA (Director)
    Comment

    I find this Declaration both timely and sound. As the director of an engineering school in computer science and applied mathematics, I would add that the same stakes apply to how we teach mathematics. Future engineers and mathematicians must still acquire, for themselves, the ability to construct and verify a proof, and the judgment to assess what automated tools could produce. If we let students delegate mathematical reasoning before they have mastered it, we erode the very understanding on which the future of the discipline depends. Teaching in the age of AI means equipping students to use these tools critically, not to depend on them.

  30. Mainak Chowdhury ORCID Member of Technical Staff, Milaneum, Inc.
    Comment

    I am delighted to see this timely initiative on protecting and celebrating mathematical precision of thought. While they may augment real-world human judgement and understanding with incomplete information, computational tools including AI are unlikely to ever replace them.

  31. Chinmoy Bhattacharjee ORCID Junior Professor, Universität Hamburg
    Comment

    This is a timely and very important declaration. As an early career researcher, I wholeheartedly agree with its statements, especially those regarding fair and ethical use of AI, its disproportional threat to early career researchers, and its potential long-term impact on what we value as mathematicians.

  32. Francesco Sica ORCID Assistant Professor, Florida Atlantic University
  33. Giada Veronelli verified email University of Milan - Bicocca
  34. Sanne ter Horst ORCID North-West University , South Africa
  35. Himanshu Sukhpal ORCID
    Comment

    We need safeguards

  36. Christopher O'BRIEN ORCID PhD Candidate, The University of Queensland
  37. Ken'ichi Ohshika ORCID Professor, Gakushuin University
  38. Tom C Brown verified email Simon Fraser University
    Comment

    Strip mining for coal (among other activities) has led us to global warming.  Peter Renz has said that "AI LLMs are intellectual strip mine operations."   (A few years ago I asked ChatGPT to "Tell me about the career of Professor X, who taught for many years at Y"    What ChatGPT told me was about 40% true and 60% nonsense.)  The recent disproof "by AI" of a long-standing conjecture of Paul Erdős is interesting - however, it took endless hours of work by human mathematicians to verify this.  Major investments in AI by governments will surely be aimed at surveillance and warfare.

  39. Bruce Rothschild Department of Mathematics, UCLA
  40. Rafael P. Saldaña, PhD Retired Associate Professor, Department of Mathematics, Ateneo de Manila University, Quezon City, Philippines
    Comment

    I stand with fellow mathematicians and scientists around the world in search for mathematical truths and ethical use of AI. Artificial intelligence should be a tool to be used to benefit humanity, not to destroy it. Integrity and excellence should not be compromised using AI.

  41. Christopher Kennedy ORCID Coleman Postdoctoral Fellow, Queen's University
  42. Gergely Zábrádi ORCID Associate professor, Eötvös Loránd University
  43. Junaid Sarfraz ORCID Associate Professor, Govt Postgraduate College No.1 Abbottabad
  44. André van Renssen ORCID Senior Lecturer, University of Sydney
  45. Geordie Williamson ORCID The University of Sydney
    Comment

    This declaration is important and timely. It is critical that our community sees the importance of affirming the importance of human understanding in mathematics (and in science more generally). Several of the other points I support unreservedly. I would like to applaud point 11 for individuals "Evaluate the ethical consequences of your work, and take action accordingly" and point 7 for organizations "Provide frameworks for collaboration". On the other hand, I think that it is already (!) unrealistic to only allow human authorship (individuals, Point 5).

  46. Sera, Martin Kyoto University of Advanced Science, Faculty of Engineering
  47. Muhammad Ismail Yunus verified email Khalifa University of Science, Technology and Research
  48. Roberto Rivera ORCID Professor, University of Puerto Rico at Mayagüez
  49. Xin Yang ORCID Ph.D. Candidate, University of Connecticut
  50. Robert M Fossum verified email University of Illinois Urbana-Champaign
  51. DINO FESTI ORCID Università degli Studi di Napoli Federico II
  52. Bach Hoang ORCID Bachelor of Science, University of Illinois Urbana-Champaign
    Comment

    Very agree with this declaration.

  53. James Newton verified email University of Oxford
  54. Shahn Majid ORCID Queen Mary University of London
  55. Eduardo Silva ORCID Postdoctoral researcher, University of Münster
  56. Constanze Liaw verified email University of Delaware
  57. Diaaeldin Taha ORCID Max Planck Institute for Mathematics in the Sciences
  58. Wolfgang Adam Heepe University of São Paulo - Business Administration; Virtual University of the State of São Paulo - Mathematics
    Comment

    Relevant debate, shall be widespread to the public. Moreover, shall stumulate further reasoning on the subject.

  59. Carlos Daniel Acosta-Medina ORCID Universidad Nacional de Colombia
  60. Jorge Estrada Hernández ORCID Assistant Professor, University of Havana
  61. Daniele Valeri ORCID Università degli Studi di Roma "La Sapienza"
  62. Ian Strachan ORCID Professor of Mathematical Physics, University of Glasgow
  63. Cruz Godar ORCID Lecturer, Yale University
  64. Emil-Alexandru Ciolan ORCID
  65. Jessica Fintzen verified email Professor, University of Bonn
  66. Grant Olney Passmore Imandra Inc and Clare Hall, University of Cambridge
  67. Stefan Weltge ORCID Technical University of Munich
  68. Martin Fränzle ORCID Professor at Carl von Ossietzky Universität Oldenburg, Germany
  69. Miltiadis Karamanlis ORCID
  70. Nicholas Rank verified email Johns Hopkins University
  71. Likun Xie ORCID Max Planck Institute for Mathematics
    Comment

    I strongly support research responsibility and transparency. Authors remain responsible for checking and standing behind their work, and AI should be regarded as a tool whose value depends on the judgment and quality of the mathematician using it. I hope this declaration can raise awareness of these issues and encourage further study and policy-making around them. AI can genuinely help mathematicians, just as other technologies have advanced science, but it may also amplify existing problems such as flawed or irresponsible papers. Another important concern is attribution. If substantial AI assistance is not disclosed, it can obscure how much intellectual or technical labor was carried out by the author and how much was aided by the tool. Mathematicians already often disclose computational assistance in checking examples, verifying cases, or supporting proofs. Similarly, AI tools should be properly described when they contribute substantively. This is especially important because AI can now assist with more than routine computation or programming: it may help identify connections between different papers, suggest proof strategies, or propose parts of an argument. These are tasks that would require technical and intellectual effort if done without AI. Therefore, when AI provides substantial assistance of this kind, such as suggesting a proof idea or pointing out a meaningful connection between different ideas, that assistance should be disclosed.

  72. Talia Shlomovich ORCID PhD in Algebra, Geometry and Mathematical Physics, Heriot-Watt University
  73. Karen Leung verified email University of Guelph
    Comment

    It's important to establish these guidelines early in the face of ever changing landscape of AI technology. Using AI in a responsible manner that adhere to the core mathematical research values and maintain critical thinking within each individual.

  74. Arne Mertens ORCID postdoctoral researcher, University of Antwerp
  75. Elia Immanuel Auer verified email Freie Universität Berlin
  76. Ahmad Abdi ORCID Mathematics, London School of Economics Department of Economics
  77. Richard Lang ORCID Researcher, Polytechnic University of Catalonia
  78. Joel Lewis ORCID Associate Professor of Mathematics, The George Washington University
  79. Marcin Magdziarz ORCID professor, Wroclaw University of Science and Technology
  80. Heikichi Hayashi verified email Adrasteia Labs
  81. Margaret Bayer ORCID Professor Emerita, University of Kansas
  82. Christopher Adam Hass ORCID Postdoctoral Associate, Rutgers University
  83. Amadeus Maldonado ORCID Northwestern University
  84. Pascal J. Kwanten Theoretical Physics Utrecht University
  85. Alex Blumenthal ORCID Associate Professor, Georgia Institute of Technology
  86. Willoughby Seago University of Glasgow
  87. Florian Pfender ORCID Professor, University of Colorado Denver
  88. Abdel-Hameed Nawar ORCID Cairo University
    Comment

    I endorse the Leiden Declaration on Artificial Intelligence and Mathematics. The principles of human ownership of results, responsibility for correctness, and transparent disclosure of tool use as articulated here are not unique to mathematics. They are foundational to all empirical and theoretical inquiry. As an economics and Political Science scholar , I see the same risks in our disciplines. Automated tools can generate regressions, models, and text that appear authoritative while misattributing sources, misinterpreting effect sizes, or substituting statistical significance for substantive significance. The Leiden Declaration’s insistence that “credit and responsibility continue to belong to humans” and that authors “retain the responsibility for correctness” should apply equally to causal claims, data analysis, and policy recommendations in the social sciences. The distinction between banning AI and assigning liability is critical. The “should we allow it” debate is indeed useless. AI will continue to be used. The question is whether we preserve standards of verification, attribution, and human judgment. Leiden Declaration answers that question loud and clear. I urge the American Economic Association (AEA), the American Political Science Association (APSA), and related bodies to adopt parallel standards. Our fields face the same pressures of fast publication cycles, proprietary models trained on our literature, and the temptation to outsource verification. The Leiden model promoting "use AI, but own the output" protects the integrity of research without stifling innovation.

  89. Kevin Kühn ORCID TU Berlin
  90. Jonathan Kirby ORCID University of East Anglia
  91. James Taylor ORCID Postdoc, University of Padua
  92. Theresa Ortscheidt verified email University of Glasgow
  93. Kasturi Barkataki ORCID Postdoctoral Research Associate, University of Edinburgh
  94. Daniele Bartolucci University of Rome "Tor Vergata"
  95. Matías Pavez-Signé verified email Universidad de Chile
  96. Rebekka Strathausen verified email Technische Universität Darmstadt
  97. Timothy De Deyn ORCID postdoctoral fellow, Max Planck Institute for Mathematics
  98. Ildefonso Castro ORCID Catedrático de Universidad / Professor, Universidad de Jaén
  99. Thomas S. Wilkinson ORCID Postgraduate Research Student, University of Exeter
  100. Ivan Hasenohr ORCID Post-doctoral fellow, Alpen-Adria-Universität Klagenfurt