AI’s New Role in Solving Math’s Toughest Problems

Artificial intelligence has started solving math problems that puzzled experts for decades. OpenAI and Google DeepMind recently made breakthroughs many did not expect so soon.

On May 20, 2026, OpenAI announced it disproved a conjecture made by Paul Erdős in 1946. This was a huge moment in math history. Jeremy Avigad, a philosophy and math professor, called it “a considerable mathematical achievement.”

At the same time, DeepMind solved nine problems, including two that remained open for 50 years. Demis Hassabis, DeepMind’s founder, said today’s AI is “extremely far from what would be a true invention or someone like Ramanujan.” Even so, these results show AI can handle complex math tasks.

The Rising Wave of AI-Related Research

The impact of AI on research papers is visible too. Top journals have seen a 20-30% increase in submissions suspected to be AI-related. Sam Livingstone from University College London pointed out that “the majority of those additional submissions are not considered good studies.” This flood of low-quality work creates challenges for reviewers and editors.

Javier Gómez Serrano, a professor at Brown University, warned, “Someone who has the tool without the training would produce trash they won’t even be able to detect.” This means just using AI tools isn’t enough. Proper knowledge is critical to avoid errors or meaningless results.

What AI Means for Human Mathematicians

Many wonder if AI will replace mathematicians. Petra Schwer of Heidelberg University said, “There will continue to be a place for mathematicians.” The human role remains vital.

Thomas Bloom from the University of Manchester added, “Recently, we have seen some new solutions with AI, but it’s still not a huge leap.” AI pushes the boundary but doesn’t yet redefine what math is.

Seewoo Lee at UC Berkeley noted, “AI helps mathematics progress more quickly. For mathematicians, that’s not always a comfortable advance.” Change can be unsettling, especially when it alters how work gets done.

Avigad also emphasized math’s deeper meaning. He said, “Mathematics is not just a game; it is an important part of how we give meaning to the world.” AI can assist, but humans bring understanding and insight.

Gómez Serrano stressed the need to adapt. “It’s important to develop competencies related to modern mathematics in the AI era.” Learning how to work with AI tools wisely will shape the future of math.

AI’s growing power is changing the landscape of mathematics. It solves old problems and speeds up research. But it can’t replace the training, intuition, and creativity of human mathematicians. The challenge now is to find the right balance between machine help and human skill.