OpenAI establishes group of mathematicians to guide AI use and research
Read more
Olhar Digital
olhardigital.com.br

OpenAI establishes group of mathematicians to guide AI use and research

OpenAI has established an independent panel composed of mathematicians with the aim of guiding the company itself and other artificial intelligence (AI) corporations regarding the intersection between AI and mathematical research.

This initiative was motivated by a series of controversies related to the presentation of mathematical results generated by OpenAI's models, as well as criticism from experts concerning issues of authorship, communication, and academic standards.

Structure and Functioning of the Group

The group, named the Advisory Group on Mathematics and Artificial Intelligence (AGMAI), will have nine members and will be based at the Institute for Advanced Study (IAS) in Princeton, USA. The researchers represent various renowned institutions, including Stanford, Harvard, Oxford, and Cambridge. Participants include mathematicians who have received the Fields Medal and those who have received MacArthur 'genius' grants.

OpenAI assures that this panel will operate autonomously, capable of issuing recommendations even without a formal request from the company. Furthermore, members have the prerogative to publicly comment on the impact of AI on mathematics and disseminate their suggestions.

According to the company, the panel's responsibilities include participating in the analysis and communication of new mathematical discoveries. Their tasks cover evaluating the relevance of findings, advising on how they should be disclosed, and proposing guidelines related to professional and academic standards in the field.

Additionally, the panel must advise on ways in which AI tools can assist both mathematical research and the learning process in mathematics. It is important to note that the members will not receive remuneration from OpenAI and have the freedom to modify the group's composition as they deem necessary.

OpenAI emphasizes that the success of this initiative depends on the researchers' ability to exercise their own judgment and to challenge the company's decisions.

Perspectives and Criticisms from Members

Martin Hairer, one of the group's members, detailed the origin of the initiative in an article published on the Proofs and Prompts blog, addressing what he classified as an 'avalanche of disinformation' about its constitution. Hairer confirmed that OpenAI initiated contact but stressed that not all members were approached by the organization. He argued that the structure is 'genuinely independent' of both OpenAI and other leading AI companies.

Hairer clarified that, except for basic confidentiality obligations, the members did not sign contracts limiting their public statements, nor do they receive financial compensation from OpenAI, relying only on technical support from the Institute for Advanced Study.

The panel's main focus, according to Hairer, is to best represent the interests of the mathematical community, although he admitted that the nine members do not form a representative sample of the entire field.

The small size of the group raised doubts about its suitability to consolidate enough knowledge to judge results in different areas of mathematics, given that cutting-edge research demands high specialization.

External Views on Representativeness

Francesco Fournier-Facio, a future mathematics professor at Heriot-Watt University, expressed ambivalence about the panel. While seeing potential in closer proximity between AI companies and mathematicians to improve the situation, he questioned whether the members truly mirror the community affected by technological advancement, describing the situation as 'like an ivory tower.'

Machado also expressed skepticism about the members' ability to represent the concerns of mathematicians who do not belong to this circle of high prestige. He commented that despite being notable mathematicians, he is unsure if they are the ideal people to represent him on political matters, as the reality of such privileged researchers may differ greatly from that experienced by most mathematicians.

Immediate Challenges and Context of Controversies

A primary challenge has already been established: upon announcing the panel, OpenAI revealed that a novel model, supposedly capable of solving the Navier-Stokes Millennium Prize problem, had solved over one hundred long-standing open mathematical problems across various disciplines. The company also mentioned having made progress in solving another Millennium Prize problem.

As indicated on the AGMAI website, managing this vast volume of results is the group's initial priority. The researchers face the specific challenge of advising OpenAI on how to coordinate the disclosure of a large quantity of significant mathematical discoveries produced by the company's internal model.

This creation follows months of criticism regarding OpenAI's methodology for presenting discoveries. Controversies included allegations of misappropriation of researchers' work, failure to give due credit to human contributions, and improper use of research conducted by individuals using their own models.

These incidents led mathematicians to question whether OpenAI possessed or respected mathematical practices and the mathematical community. Hairer argued that it would be inconsistent not to dialogue with the company at this time, given its apparent effort to establish a link with researchers. Although the group has not yet defined recommendations, it has already identified the central issue to be addressed.

He observed that in recent months, AI companies have generated mathematically impactful results, but the way these discoveries were published and disseminated fell short of what would be considered acceptable mathematical practice. Currently, the group is gathering opinions through conversations with peers, a public contribution form, and debates on Proofs and Prompts, which has become a discussion forum on the topic.

Hairer also warned about the communicative power of AI companies, considering it naive to believe that laboratories would not try to present the panel's statements in a manner favorable to their own public relations strategies.

Kevin Buzzard, a mathematics professor at Imperial College London, questioned the need for such an exclusive group for OpenAI to understand what the mathematical community desires. He stated that 'these were not proofs of difficult theorems, but a better understanding of our field,' adding that 'it is not entirely clear to me that you need a committee of brilliant people to emphasize this point.'

Similar stories

OpenAI announces solution to one of the Millennium Prize Problems
Read more
podrobno.uz

OpenAI announces solution to one of the Millennium Prize Problems

OpenAI has announced that its internal artificial intelligence system was able to find a solution to one of the seven Millennium Prize Problems—the Navier–Stokes existence and smoothness problem. Mathematicians have been working on this issue for many decades.

Along with the statement of the proof, the company provided its formalized version in the Lean system. The Navier–Stokes equations are used to describe the movement of gases and liquids, for example, when modeling airflow, weather, and blood circulation. The main point of contention concerned the possibility of singularities arising—states that violate the smoothness of the mathematical model—during the initially smooth movement of a three-dimensional fluid over a certain period of time. OpenAI claims that its system demonstrated such a possibility.

According to the company's data, approximately 10,000 AI agents participated in this task. The solution was reached on September 5th after about 88 hours of operation, and the subsequent formalization and verification of the proof in Lean required an additional 17 hours. During the process, the agents exchanged 2.7 million messages and generated approximately 130 billion output tokens.

The journal Nature noted this announcement as a potentially significant breakthrough in mathematics, but for now, it is only a claimed result. For official recognition of this achievement, the Clay Mathematics Institute requires that the proof undergo independent review and receive broad approval from the scientific community. According to the institute's regulations, the candidate for the prize must publish the solution at least two years before consideration. A prize of one million dollars is allocated for each of the seven Millennium Prize Problems, but OpenAI stated its refusal to claim this award.

Furthermore, it was previously reported that the GPT-6 Astra model successfully passed all 48 levels of the 'I am not a robot' test, which simulates various types of CAPTCHAs. To complete these tasks, the artificial intelligence independently interacted with the interface and performed screen content recognition.

Popular