OpenAI has once again drawn criticism from the mathematical community, this time over its handling of a newly formed independent advisory group intended to repair relations with researchers. In recent months, the company has produced notable breakthroughs in mathematics, yet its announcements have repeatedly been met with confusion and frustration among practitioners.
The latest effort centres on consulting an independent advisory group of elite mathematicians. However, mathematicians speaking about the matter, including one member of the group itself, have described a messy and confusing affair bearing many of the hallmarks of the company’s previous rushed forays into the field. The situation suggests that little has changed despite earlier missteps.
The Formation of AGMAI
On 21 September, a group of eminent mathematicians announced the creation of the Advisory Group on Mathematics and Artificial Intelligence (AGMAI). The announcement appeared as a guest post on the blog of UCLA mathematics professor Terence Tao, a Fields Medalist and an outspoken critic of AI companies’ conduct in mathematics.
The body was intended to comprise nine elite mathematicians operating at the very top of their field. According to the post, the group would function independently from OpenAI and advise the company, along with other frontier AI labs, on the review and communication of emerging results. In its own statement, OpenAI repeatedly emphasised the group’s independence, noting that members would be free to challenge the company publicly, publish their advice, and offer guidance that had not been requested. The company also stressed the limits of the group’s influence, stating that it would not be responsible for advising on how the firm paces its internal progress in mathematics.
Confusion Over Independence
Many mathematicians, and most people outside the community, learned of the group primarily through OpenAI’s louder public rollout. The result was widespread uncertainty over whether AGMAI was genuinely independent or an OpenAI-appointed body. That impression was not unreasonable: the AGMAI website states the group formed after OpenAI approached some of its eventual members about establishing an advisory board, before those individuals decided to strike out independently and invite others to join. It remains unclear which of the nine members the company initially approached.
Martin Hairer, an AGMAI member and mathematics professor at Imperial College London and the Swiss Federal Institute of Technology in Lausanne (EPFL), was keen to stress the group’s independence. While acknowledging that OpenAI’s approach to some mathematicians was the original impetus for its formation, he said the group receives no financial, technical, or other support from the company. He added that none of its members have signed agreements restricting what they can say or do beyond standard confidentiality requirements needed to grant them advanced access to research.
The group’s first task is to help coordinate the release of numerous further results that OpenAI says its unreleased model has produced, a prospect already stirring concern among researchers over the potential impact on their field.
Source
Image: theverge.com
