How can artificial intelligence help researchers process historical documents? Could AI solve some of mathematics’ hardest open problems? And what happens when technological progress moves faster than the development of adequate safety measures? These are among the questions explored at Researchers’ Night on 25 September at Alfréd Rényi Institute of Mathematics in Budapest, where this year’s programmes share a common focus: the rapidly changing relationship between artificial intelligence, mathematics and research.
The Rényi Institute will present two very different perspectives on AI. One programme, organised jointly with the Historical Archives of the Hungarian State Security Services (ÁBTL), looks at the digitisation and AI-assisted processing of the archives of Hungary’s former state-security apparatus. The other asks a provocative question at the frontier of mathematics and AI: What if an artificial intelligence solved one of the seven famous Millennium Prize Problems – and who would get the million-dollar prize?
1. AI and the archives of the state-security services
Digitising historical documents is no longer simply a matter of scanning paper records and making them searchable. AI offers new ways of processing and connecting vast amounts of complex archival material, potentially opening previously inaccessible sources to researchers.But this also raises important questions. How can AI be used without losing the historical context of a document? What can current AI tools actually do, and where are their limitations? And what might be the social consequences of making previously difficult-to-research historical records accessible in new ways?
Following three short ten-minute presentations, historians, archivists and mathematicians will discuss these issues in a roundtable organised by Rényi Institute and ÁBTL. Participants include Gergő Bendegúz Cseh, Director-General of ÁBTL; Eszter Tóth, Deputy Director-General of ÁBTL; Dezső Miklós, Deputy Director-General of Rényi Institute; Gábor Kovács, mathematician at Rényi Institute’s Department of Artificial Intelligence; and Sándor Horváth, Head of the Contemporary History Department at ELTE's Research Institute of History.
The roundtable begins at 4:00 pm. Registration is required via the Researchers’ Night website (Sry, it is in Hungarian, ed.)
2. Can AI solve a million-dollar mathematics problem?
The second programme approaches AI from a completely different direction: the world of the Clay Mathematics Institute’s Millennium Prize Problems.In 2000, the Clay Mathematics Institute selected seven exceptionally difficult mathematical problems and offered a prize of $1 million for the solution of each. The problems include the Riemann Hypothesis, the P versus NP problem, and the Navier–Stokes problem concerning the mathematical description of fluid motion. So far, only one of the seven, the Poincaré Conjecture, has been officially solved. Grigori Perelman’s solution was subsequently verified by the mathematical community.
The story has taken a remarkable new turn this month. On 8 September, OpenAI announced that an internal AI system had found a solution to the Navier–Stokes problem. The claimed proof is still being scrutinised by independent mathematicians, and the announcement has already sparked debate within the mathematical community. Read Quanta's article HERE.
This makes the question at the heart of the Rényi programme particularly timely: What happens if an AI really does solve a million-dollar mathematical problem? Who owns the proof – the machine, its developers, or the mathematicians who verify the result and turn it into a comprehensible mathematical argument? Can an AI-generated result be considered a mathematical discovery? And if AI becomes capable of solving problems that have challenged mathematicians for decades, what will the role of human mathematicians be?
These questions will be debated by András Stipsicz, Director-General of Rényi Institute, and Tamás Titkos, mathematician and researcher at the Institute, in an “angel versus devil” style debate beginning at around 5:30 pm. No registration is required.
The debate is also part of a much wider and increasingly urgent discussion about the pace of AI development. In recent days, leading figures in the AI industry have called for development to slow down so that safety measures can catch up with rapidly advancing technology.
For Rényi Institute, Researchers’ Night is an opportunity to bring visitors closer to questions at the forefront of mathematics and related sciences. This year’s programmes highlight a particularly striking development: mathematics is not only one of the foundations of artificial intelligence – it is itself becoming a field that AI may fundamentally transform.
The full Researchers’ Night programme at Rényi is available through the official event website.