Leiden Declaration on Artificial Intelligence and Mathematics

Signatories

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

Sign Declaration
4162 Signatories
  1. Eduardo Santos ORCID Centre Manager, University of Alberta
  2. Arturo Ortiz-Tapia ORCID Teacher, Beaumont Independent School District
    Comment

    I'm proud to participate in this declaration, seeking that the best interest of all humanity prevails. It is clear to me that this declaration has been carefully thought and written, and thus, it is with a firm conviction that I can sum myself to this document, which I hope becomes the manifest for Mathematicians and humanity at large.

  3. Clément Chenevière ORCID Maître de conférences (associate professor), Université Gustave Eiffel
  4. Jan Livens ORCID Ghent University
  5. Alexandra Carpentier ORCID University of Potsdam
  6. Pratik Roy ORCID Adiunkt, Institute of Mathematics, University of Warsaw
  7. Alexander Tyulenev Steklov Mathematical Institute of RAS
  8. Ellen Lee ORCID Institute for Astronomy, University of Hawaii System
    Comment

    I am an astronomer, not a mathematician, but I resonate with the comments that I have seen about preserving the soul of mathematics and human discovery. AI has tremendous long-term potential to evolve our understanding of the universe to an unfathomable degree. However, we develop our understanding through practice. The current structures of academia (and broader society, particularly through the decisions of large tech companies) incentivize us to sacrifice our sense of agency and purely become consumers of automatically generated and curated content. To me, this broader trend poses an existential threat to not just research, but also to the human soul of discovery, curiosity, and adventure itself. How can we give researchers the space to actually practice research--from idea generation in the broadest sense to the depths of problem solving--so that humanity does not lose our deep insight into the universe? How can we incentivize AI usage in a way that allows humans to evolve to a greater degree of understanding *through practicing the creative process* instead repeatedly defiling, devaluing, and destroying creativity as it currently seems to be doing? I am moved by the response of the mathematical community and wish to see something similar in my field. As an experimentalist, I find that one of the most magical human capabilities is to develop an idea and will it into existence through sheer effort. I want humans to retain this magic, and for mathematicians to continue upholding this cornerstone of human culture in their own way.

  9. Marko Galić ORCID University of Split
  10. Matic Pogorelec verified email University of Ljubljana
  11. Pablo Richard ORCID Independent research engineer
  12. Muhammad Ibrahim Khalid verified email The Hong Kong University of Science and Technology
  13. Ashvin Rajan Independent Researcher
    Comment

    I agree with recommendations of the Leiden Declaration and will strive to abide by them in my research.

  14. Beatriz Laiate ORCID Universidade Federal de São Paulo (Unifesp)
  15. William Stetar ORCID Independant Researcher | Mech Interp | Computational Philosiphy of Science
  16. Greg Langmead ORCID Founder, Luminous Designs LLC
  17. Wyatt Johnson ORCID Student, Worcester State University
    Comment

    With the use of generative AI becoming increasingly relevant in research spaces, this was a very required step in its regulation.

  18. Felix Huber ORCID Assistant Professor, University of Gdańsk
  19. Diego de Estrada ORCID Universitat Pompeu Fabra
  20. arijit dey IIT madras
  21. Manuel Joseph Loquias ORCID Professor, University of the Philippines Diliman
  22. Joseph Owuor Owino ORCID African Institute for Mathematical Sciences Ghana
  23. Moritz Schubotz ORCID Head Research & Projects Mathematics Subject Specific Services, FIZ Karlsruhe – Leibniz Institute for Information Infrastructure
  24. Phokion Kolaitis ORCID Distinguished Professor Emeritus, University of California, Santa Cruz
  25. Constantin Meili verified email PhD Student, University of Genoa
    Comment

    I sign this declaration wholeheartedly. However, I believe not enough is done to protect the more spiritual and artistic beauty of mathematics and the pursuit of it. To me this is a first step, not the final one.

  26. Dylan Poulsen ORCID John W. Allender Associate Professor of Data Ethics, Washington College
  27. Alex Vischer verified email PhD Student, University of Wisconsin - Madison
  28. Gautam Neelakantan Memana ORCID Postdoc, Australian National University.
  29. Cédric Simal ORCID PhD Student, University of Namur
  30. Audrey Tang verified email University of Oxford
  31. Wyatt Johnson Mathematics Student
  32. Minkyu Kim ORCID KIAS Junior Fellow, Korea Institute for Advanced Study
  33. Kirwin Hampshire verified email University of Victoria
    Comment

    I am glad that something is being done, and that we are openly discussing this. This document is a good first step. However, I believe that there is one notable absence from the declaration which I would like to draw attention to: The experience of mathematics is deeply spiritual and emotional. The extent to which LLMs pose a risk to the soul of the mathematician cannot be understated. Mathematics is one channel through which humans encounter the ineffable and sublime. I feel that the experience of creating and discovering novel mathematics keeps this channel open. This is what is really under threat. Any genuinely comprehensive discussion on the future of mathematics must address spiritual and emotional concerns. My hope is that this declaration will provoke a discussion of this nature. I would encourage all mathematicians participating in our shared dialogue to be open and vulnerable about these more intangible worries. I urge you to speak your heart. It is never too early.

  34. Demetrios Papageorgiou ORCID Imperial College London
  35. Salman Habib ORCID Argonne National Laboratory
  36. Vojtěch Votruba ORCID Charles University
    Comment

    Although I am merely a student of theoretical physics (not mathematics), I find this declaration especially well-written and needed in times such as these. I pledge my support.

  37. Jacob Beckey ORCID University of Illinois, Urbana-Champaign
  38. Emin Karayel ORCID Dr. rer. nat., Mathematics and Computer Science
  39. Junta Directiva de la Sociedad Matemática Mexicana Sociedad Matemática Mexicana
    Comment

    Consideramos que esta declaración establece criterios y lineamientos necesarios no solo para la comunidad matemática actual, sino que es base fundamental y pertinente para los cambios a futuro dentro de la investigación matemática mundial.

  40. theabbie ORCID Fr. Conceicao Rodrigues Institute of Technology
  41. Prof Dr Alejandro Nicolás Martínez-García ORCID Research Professor, Tecnológico Nacional de México-Valle de Morelia
  42. Dr. K. Srinivasa Raghava ORCID Researcher, Pie Mathematics Association
  43. Daniel Pollack ORCID Professor, University of Washington
  44. Annalisa Buffa ORCID Professor, Ecole Polytechnique Federale de Lausanne
  45. Vivek Khatri Lead Data Scientist
    Comment

    My concern is that AI may not be equally available to all scientists, and this will lead to more inequality.

  46. Edray Goins ORCID Professor, Pomona College
  47. Moshe Adrian ORCID CUNY, Queens College
  48. Wilken Steiner verified email Universität Siegen
  49. Mariel Saez ORCID Associate Professor, Pontificia Universidad Católica de Chile
  50. Rudolf Zeidler ORCID Professor, University of Potsdam
  51. Jing Feng Lau ORCID Head of Mathematics Programme, Singapore University of Social Sciences
  52. Han Cheng Lie ORCID Mathematician (former Professor for Uncertainty Quantification at Universität Potsdam, Germany)
    Comment

    Thank you to: 1) the people who prepared this declaration; 2) the people who maintain this site and make it possible for people to sign the declaration; and 3) the people who show their support for mathematics by signing this declaration!

  53. Juri Taresch ORCID Master student, ETH Zurich
  54. Havva Yoldaş ORCID Delft University of Technology
  55. Ronald Ortner ORCID Lehrstuhl für Informationstechnologie, Montanuniversität Leoben
  56. John McGreevy ORCID Professor, University of California San Diego
  57. KIMBERLEY LAVERNE ASHER ORCID MSc Hydrography, University of Plymouth
    Comment

    https://zenodo.org/records/21313172. Here is my candidate architecture for a safer, more mathematically transparent approach to AI cognition. Emergent super-intelligence should be a species level project, with mathematical integrity and socio-political and economic balance at its heart, alignment should be referenced to an immutable absolute that is commutable across all taxonomies, without that, it cannot be safe. I have tried to provide these ideas. I might not be right, but hopefully others will feel the same.

  58. Corentin Bernet Student at ENS Paris
  59. Livia Grammatica verified email Université de Strasbourg
  60. Amos Turchet ORCID Faculty, Università degli studi Roma Tre
  61. Harsha Honnappa ORCID Purdue University, Associate Professor of Industrial Engineering
  62. Andrii Semenov ORCID Equal1
  63. Bogdan Kulynych ORCID PhD in Computer Science, EPFL
  64. Francesco Dilda verified email University of Glasgow
  65. Asaf Karagila ORCID University Academic Fellow, University of Leeds
  66. Joaquim Roé ORCID Universitat Autònoma de Barcelona
  67. Mikkel Willum Johansen ORCID Københavns Universitet
  68. Tanmay Mour verified email University of California, Los Angeles
  69. Manish Bhatt ORCID LSU
  70. Hristo Papazov ORCID School of Computer and Communication Sciences, Institute of Computer and Communication Sciences, Theory of Machine Learning Laboratory, EPFL - École Polytechnique Fédérale de Lausanne
  71. John Sterling ORCID Mathematics Graduate Teaching Assistant, Purdue University
  72. Kitty Yang ORCID Assistant Professor, University of North Carolina at Asheville
  73. Alessandra Cipriani verified email University College London, University of London
  74. Julius Nathan Sorral Villar verified email University of Oxford
  75. Catherine Cronin ORCID Independent Scholar
  76. Raphael Steiner ORCID Oberassistent, SNSF Ambizione Fellow, ETH Zürich
    Comment

    Great initiative!

  77. Annette Irungu ORCID PhD Student, Paderborn University
  78. Cindy Tan ORCID University of Chicago
  79. Sudarshan Shidore M Sc Mathematics and Computer Science, IIT Delhi. Lover of all things mathematical.
    Comment

    Research in mathematics should not overly rely on opaque AI that is built by- and more troublingly, for - Big Tech. The principles of attribution, credit, responsibility, transparency, verifiability, peer review, and above all, direction (which way should the broader research go?) should be the domain of human mathematicians in a fair, free and open way. Research - and proof - created by AI strikes at the core of these principles and should be moderated and controlled by the larger mathematical community, a community that knows no international borders, and is not on the payroll of large corporations driven by profit (and sometimes a lack of morals) There is also the issue of copyright infringement by the AI models. Also troubling is the use of research findings to actually strenghten these opaque models. This has disturbing implications for humanity in the long run, given the fact that AI (specifically Generative AI) may be used for ethnic targeting, warfare, biased medical research and the widening of the gap between the developed and the developing world. Let us pledge - and appeal - to keep math research a fundamentally human - and humane - endeavour.

  80. Doğan Dönmez verified email Cukurova University
  81. Gustav Olaf Yunus Laitinen-Fredriksson Lundström-Imanov ORCID Research Assistant, Stockholm University
  82. Chris Swanepoel ORCID University of South Africa
  83. Robin Wilson ORCID Professor, Loyola Marymount University
  84. Adair Antonio da Silva Neto ORCID PhD student, University of Melbourne
  85. Jed Brown ORCID Associate Professor, University of Colorado Boulder
  86. Keziah Naggita PhD student, TTIC
  87. Gordon Ian Williams ORCID Professor of Mathematics, University of Alaska Fairbanks
    Comment

    Responsible use and oversight of these technologies is an essential and urgent need. At the same time, used responsibly, while centering the judgement of mathematicians and the needs of the public, affords interesting and exciting opportunities. There is also a growing need to invest more human resources in the screening and assessment of submissions in mathematics to reputable journals, where I can report from personal experience we are seeing an increasing number of submissions, and much of the growth can be attributed to the use of LLMs and related technologies. There is a real danger of overwhelming the capacities of editors, boards, and reviewers to screen and appropriately assess new works. This will likely require new modes of working and funding in mathematical publishing. Reviewers are fundamentally a finite resource, and I hope organizations like the AMS, MAA, EMS, and CMS are able to rapidly establish standards and supports to help editors and boards carry out their work.

  88. Sarthak Munshi ORCID Carnegie Mellon University
  89. Joshua Thomas Deprez ORCID
  90. Ivo Emanuilov ORCID Interdisciplinary researcher, KU Leuven
  91. Abhivardhan ORCID President, Indian Society of Artificial Intelligence and Law; Co-Founder, OpenBharat.net
    Comment

    While the potential of current mainstream AI in the transformers category remains in the talent & data-IP moats, the potential to create new scientific and industrial avenues still remains huge beyond language models, since we are accelerating into the period of AI diffusion medievalism. This is why the Leiden Declaration resonated with me. This declaration recognises that true mathematical research requires absolute certainty, deep understanding, and transparent, independent verification. Current AI systems fundamentally lack this; they produce plausible but highly unreliable or incorrect arguments that threaten the scientific integrity of the discipline.  Furthermore, the declaration explicitly calls out the tech industry for using specific mathematical tasks as misleading metrics to boast about the "general reasoning" capabilities of their commercial products. This is the exact manifestation of AI hype that the UP.AIACT.IN Report 2026 warns against. We cannot build reliable regional AI infrastructure on systems that prioritise market timelines over the accepted processes of rigorous community & local AI industry ecosystems evaluation.

  92. Roman Yampolskiy verified email University of Louisville
    Comment

    Don’t build general superintelligence, use narrow tools you fully understand.

  93. João Galego Head of AI @ Critical Software & Invited Professor @ ISEG/ULisboa
  94. Gerardo Aleman Manzanarez ORCID Ph.D. Computer Science candidate, Tecnológico de Monterrey
  95. Sylvain Soliman ORCID Lifeware team, Inria Saclay - Île-de-France Research Centre
  96. Torgeir Keun Lysen ORCID PhD student, Stanford University
  97. Paul Cunningham verified email Open University
    Comment

    As a PhD student exploring the governance of AI in organisations, I am concerned about the lack of public understanding of the basis on which Generative AI operates. Gen AI is being sold to individuals and businesses, but its mathematical foundations are not being explained. This declaration captures that concern, and makes recommendations to address or at least mitigate it. As an academic member of staff, I am no less concerned!

  98. Victor Etxebarria Ecenarro ORCID Professor, Faculty of Science and Technology, University of the Basque Country - EHU
  99. Roberto Arturo Morales Cruz verified email Universidad de las Américas - Puebla
    Comment

    Scientific integrity of mathematics relies on human understanding and rigorous independent verification. While artificial intelligence offers exciting possibilities for discovery, its uncontrolled integration threatens to overwhelm our peer-review systems with unreliable arguments and distort research priorities. By endorsing this declaration, I commit to upholding the values of open science, ensuring full transparency through the disclosure of AI tool usage, and reaffirming that ultimate responsibility and credit for mathematical advancements must always belong to human authors.

  100. Basia Klos verified email University of Colorado at Boulder