Mathematicians urge halt to OpenAI collaboration over bulk math problem releases

The release of solutions to 372 long-standing mathematical problems by OpenAI has triggered a significant backlash within the academic community. According to TechRadar, renowned mathematician Terence Tao has called for colleagues to abandon projects involving the company. This reaction stems from concerns that the scale and complexity of the output undermine traditional scientific values centered on human understanding.

What did OpenAI publish and how was it generated?

OpenAI published solutions or partial solutions to 372 outstanding mathematical challenges that have puzzled researchers for decades. These outputs were generated using an internal, unreleased specialized AI model. The company released the work across 722 papers, a volume that TechRadar notes makes it difficult for human mathematicians to verify or fully comprehend the methods used.

Why are experts criticizing this release?

Leading experts argue that the release demonstrates power rather than scholarship. The Association for Human Mathematics stated that mathematicians did not request this work. They highlighted that the Advisory Group on Mathematics and Artificial Intelligence, which OpenAI cited for legitimacy, previously advised against testing advanced problems on internal models. The group urges a return to science that prioritizes human comprehension over automated output.

How does this compare to recent AI capabilities?

TechRadar reports that just two years ago, AI struggled with GCSE-level mathematical challenges. Now, the technology is addressing problems that took humans years to establish. OpenAI recently published a proposal for the Navier-Stokes equation using its unreleased model. The rapid shift from basic arithmetic struggles to solving iconic, complex problems has stunned the advanced mathematics community.

What are the implications for organizations?

For organizations relying on AI for complex problem-solving, this event highlights the gap between machine output and human verification. The Association for Human Mathematics warns that releasing over 700 files at once is not a demonstration of scholarship. Businesses must consider whether automated solutions can be understood by their technical staff. If the underlying logic cannot be verified by humans, the operational risk of adopting such outputs remains high.

What to do and how to stay safe: OpenAI

• Review internal policies on using AI-generated outputs for critical mathematical or scientific calculations.
• Ensure that any AI-assisted research includes a mandatory human verification step before publication or implementation.
• Monitor communications from professional associations regarding ethical guidelines for AI collaboration in technical fields.
• Assess whether current AI tools provide sufficient transparency for your organization to understand the reasoning behind complex results.

Step-by-step guide: AI Usage Policy Checklist for Secure Enterprise Deployment

General security guidance from the Sandbox Desk newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

How many mathematical problems did OpenAI solve?

OpenAI published solutions or partial solutions to 372 outstanding mathematical problems.

Who is leading the opposition to OpenAI's release?

Renowned mathematician Terence Tao is leading the opposition, supported by the Association for Human Mathematics.

What is the Association for Human Mathematics asking mathematicians to do?

They urge mathematicians to discontinue their work with OpenAI and return to a vision of science centered on human understanding.