Rényi's Researchers Make a Meaningful Contribution to Science Communication

In October, Rényi's researchers discussed the use and impact of artificial intelligence in a number of in-depth interviews.

The statements made by researchers at Rényi Institute paint a clear picture of artificial intelligence as both a scientific opportunity and a serious challenge, including its social consequences. The Institute’s researchers do not see AI simply as a new and useful technological tool. They describe a transformation that could reshape the methods of mathematical research, the acquisition of knowledge, the role of human creativity, and even our understanding of self.

The views expressed by Dániel Varga and Ákos Dúcz on the use of artificial intelligence in mathematics illustrate how the new technology can be put to work in service of science, potentially helping solve problems that once seemed inaccessible. The two researchers found a solution to part of one of Paul Erdős’s problems, which had remained open for decades, using AI primarily as an auxiliary tool. They relied on it for programming tasks and computations, significantly accelerating their research: they were able to test ideas more quickly, while an unexpected response from AI turned out to be an important step towards the solution. However, the mathematical insight and the choice of research direction remained the work of human researchers. The interview was published by Qubit and its full version can be read in English full HERE in our related article.

According to Dániel Varga, “fantastic results can still be achieved in mathematics after the advent of artificial intelligence”, although he is less certain that mathematicians themselves will adapt easily to the changes ahead. Researchers will face an increasing challenge in identifying where human input can add value to machine performance and how the value of mathematical work will change when AI systems begin to contribute to solving problems that were previously considered extremely difficult. Ákos Dúcz observed that he can currently work effectively with AI while remaining in control of the process, but that this situation could change rapidly. Dániel Varga’s remarks suggest that artificial intelligence may already be capable of more than simply assisting with routine tasks: it can also lead to unexpected insights that open up new avenues of research.

Balázs Szegedy, academician and head of the Artificial Intelligence Research Department at Rényi Institute, believes that artificial intelligence could transform not only science but also people who practise science. He added that AI is the first human invention capable of producing new results itself and competing with human abilities in creative and intellectual activities. In the past, computers were primarily effective at precise, well-defined tasks, while creativity was considered an exclusively human domain. His responses on the RTL Reggeli programme on 15 October 2026 can be viewed (in Hungarian) HERE.

“This could have particularly serious consequences for education,” he added. If students rely on artificial intelligence to solve assignments and teachers use it to assess their work, essential parts of the learning process may be lost. Although using the tool can help us finish tasks more quickly, we may fail to develop precisely those abilities that we will need in the future. Balázs Szegedy is convinced that AI should not be discussed solely as a technological issue: we must also address it as a social concern and understand both its potential and its limitations. Alongside the risk of job losses, he highlighted the deeper social and psychological effects of AI. Technological transformation is not merely an economic issue: it can also affect our self-esteem and our understanding of the meaning of life. Moreover, the pace of technological development exceeds the speed at which members of our societies can understand and process its consequences.

Alongside these risks, the head of the Rényi Institute’s Artificial Intelligence Research Department also sees positive opportunities. Used appropriately, AI could support collective decision-making and contribute to a more stable and equitable society.

Klubrádió's (a Hungarian Commercial Broadcast Channel) programme called Mesterséges Értelem (Artificial Sense or Intellect)) highlights another, distinctly practical application of the technology: AI can help process difficult-to-read historical documents, making our shared body of human knowledge more accessible. At the same time, human expertise remains indispensable for checking and interpreting machine-generated results. The programme, broadcast on 24 October under the title “Artificial Intelligence Helps Decipher Secret Agent Files”, can be listened to in Hungarian HERE. An article by print Hungarian weekly Szabad Föld, based on the events of the Rényi roundtable discussion on Researchers’ Night, is also available HERE.

How will the rise of artificial intelligence change the process of mathematical research, particularly the proving of theorems and conjectures? This was the question posed by the host of Felkelő (Riser), a programme on national public radio Kossuth Rádió, to Győző Csóka, an AI researcher at Rényi Institute, in an interview. He emphasised that AI can already write research papers once the right questions have been posed by humans. Like Miklós Abért, whose views are also quoted below, he believes that the advent of AI is bringing about the greatest transformation in mathematical research in the past century: language models can independently make mathematical discoveries and introduce new abstractions, fields and perspectives. “One positive aspect, however, is that you can take an old, unfinished research paper off the shelf, have AI read through it and add to it, or even turn it into a complete paper,” he added. Győző Csóka, too, sees benefits in the collaboration between humans and AI, with the latter acting as an exceptionally smart and tireless assistant to the researcher. 

Nothing comparable in significance has occurred in modern mathematics for 200 years, Miklós Abért, senior researcher at Rényi Institute’s Algebra Research Department, noted on Felfedező (Discoverer), the science programme on Hungarian PBS Kossuth Rádió, referring to OpenAI’s release of almost 400 result families. This remains true even if only some of the proofs ultimately turn out to be correct. He also pointed out that the mathematical community has not yet verified whether most solutions to these problems, which lie at the forefront of mathematical research, are valid. In three cases, AI has already been found to be wrong.

“A good proof is not the same as a complete mathematical understanding. The mathematical community must process these results and incorporate them into its own language, and that takes time. No algorithm can replace this process, because it is a collective human activity.”

Miklós Abért emphasised the importance of the paths leading to mathematical proofs and the need to reorganise the entire mathematical landscape in response to OpenAI’s surge of results. He stressed that AI represents an extraordinary engineering achievement, but is not currently a scientific entity in its own right. We still do not understand why it works or what its limitations are. “These are systems operated in pursuit of commercial interests, but that does not make them any less dangerous. Not because they might become conscious, but because of their potentially devastating effects on education and learning, for example,” Abért Miklós emphasised.

Ultimately, the views expressed by researchers at Rényi Institute point to a radical transformation in the role of scientists, and not only in mathematics. We need to rethink the nature of research, the purpose of education and the abilities worth developing in a world where an increasing number of tasks can be automated and performed on our behalf, potentially even better than we can fulfill them ourselves. Understanding technology is a shared social responsibility. It is not enough to benefit from AI: we must also understand its risks so that we do not become vulnerable or give up independent thought. It is not the future of knowledge and creativity itself that is at risk, but the way we pursue them. This makes it essential to reconsider the role of human beings in the creation of knowledge.