(I did not use generative AI to write this text (apart grammar/ proof checking). It was originally written in French and it was translated with the help of QWEN3.8, an open weight model.)
It was the image of a bombardment that came to my mind on the day OpenAI announced the resolution of 372 major problems in mathematics across 719 manuscripts: destroyed theses, shattered academic lives, research fields in ruins, and the evidence of a brutal shift of era in which a company asserts its superpower without regard for those it has just crushed.
I am not in the habit of giving my opinion publicly on my discipline. I trust collective deliberations more, which carry far more weight and legitimacy. But reading the very numerous high-quality contributions on this site, I wanted to add my small stone to this collective work of mourning our past life, of which this site is becoming, day by day, the place.
Why be so dramatic? The young people who embrace the demanding path of mathematics are our most precious collective asset. They may be the first victims of the revolution underway. A decline in positions is to be feared as early as 2027 in North American universities, and it could spread throughout the world. Students risk turning away from mathematics. Without even speaking of uncertainties in the job market, the symbolic and romantic force of our discipline has collapsed, and without it, we will no longer attract young people seeking understanding and universalism.
Obviously, the production of new knowledge in mathematics has exploded and one can only marvel at the truly superhuman feats accomplished by generative AIs. I will not dwell on the reasons pushing OpenAI, Anthropic and others toward this trophy race to theorems, we are but collateral victims of stakes that far exceed our community. Even when these companies turn away from mathematics, the AI revolution in our discipline will continue its development. It might perhaps even open up desirable horizons. But for that, we must organize ourselves to collectively survive this change of era.
Here are some proposals:
- Let us preserve the unity of our community and keep dialogue open between our different opinions and strategies in the face of this revolution. The transparency on AI tools and models used, if any, should be the norm in our works.
- Let us stop using OpenAI tools to conduct research and let us write in our articles “We did not use for this project generative AI provided by OpenAI, whose practices do not meet the standards of our community”. As highlighted by the Association for Human Mathematics, this company is out of step with the ethics of our profession and has shown us nothing but contempt, even in the face of our dialogue initiatives such as that of the Advisory Group on Mathematics and Artificial Intelligence, or as illustrated by the testimony of Tristan Buckmaster. Its practices are opaque and take advantage of our completely open academic world. I know of no real necessity for such an acceleration of mathematical research to meet the needs of other sciences or of society. Let us take the time to digest what has been produced in a few months.
Beyond the case of OpenAI, we should demand of all concerned companies that they accept the most elementary ethical recommendations, such as those formulated in the Leiden Declaration or by the Advisory Group on Mathematics and Artificial Intelligence. In a few years, we could have an open-source generative AI model controlled by the academic world. With such a model, we will have control over the financial and ecological costs we are willing to bear to advance our discipline, and we will be able to protect ourselves against a multi-speed discipline, depending on the wealth of universities and their proximity to AI companies.
- Let us fight to preserve positions, at least initially. We have trained young mathematicians: it is our duty to help them find employment, security and guarantee fair conditions of access to these jobs regardless of their use of generative AIs.
- Let us redefine the contours of our discipline and our profession. New fields and practices of mathematics will emerge, interdisciplinary research will develop, public outreach and mathematical culture could enter a golden age, mathematics becoming so accessible. This could prevent the foreseeable decline in the number of mathematics students.
- Let us invest in the field of AI safety. We are experiencing this directly: AI raises very numerous critical challenges for our societies, such as cybersecurity and financial security, or the risk of cognitive collapse. Mathematics can contribute to this emerging field. We cannot leave it to the companies developing AI to decide what is the proper use of AI for tomorrow’s society.
- Let us defend a right to non-use of AI in all professional tasks, in all professions, in the same way that we have a right to digital disconnection.
The title of this text refers to Germany Year Zero, a masterpiece by Roberto Rossellini. It follows a child, alone in a Berlin in ruins, who brings to mind our young mathematicians. I hope that we will know how to avoid a tragedy.
Received 12 October 2026.
Add to the discussion