The AI Revolution in Math Has Arrived
By Konstantin Kakaes
April 13, 2026
Introduction
The landscape of mathematics is undergoing a significant transformation, driven by advancements in artificial intelligence (AI). The tipping point for this revolution occurred in the summer of 2025, when AI models demonstrated remarkable capabilities by solving five out of six problems at the International Mathematical Olympiad, an annual challenge for some of the world’s brightest high school students. This unexpected achievement caught the attention of mathematicians and researchers, leading to a surge of interest in the potential of AI in mathematical research.
The Impact of AI on Mathematical Research
While the results from the Olympiad were impressive, they did not immediately translate to breakthroughs in open mathematical questions. However, they sparked curiosity among mathematicians who had previously dismissed AI as too error-prone. Early adopters began experimenting with AI models and discovered that these tools could assist in making genuine advancements in mathematics.
Prominent mathematician Terence Tao from the University of California, Los Angeles, remarked, “2025 was the year when AI really started being useful for many different tasks.” Although the new results may not be groundbreaking individually, they are comparable to discoveries published in professional mathematical journals. In some instances, AI algorithms were able to formulate conjectures, prove them, and verify the proofs with minimal human intervention. Additionally, extensive discussions with large language models like ChatGPT and Claude have led to novel proof strategies.
Changing the Nature of Mathematical Work
The integration of AI into mathematics is fundamentally altering how the discipline is approached. Traditionally, mathematicians focused on solving one problem at a time. However, with the advent of AI tools, they can now tackle thousands of problems simultaneously and conduct statistical studies on a larger scale. Tao noted that while AI will not replace mathematicians, it will necessitate significant cultural and institutional changes within the field.
Concerns About AI’s Role in Mathematics
Despite the excitement surrounding AI, some mathematicians express concerns about its potential impact. Akshay Venkatesh, a Fields Medal recipient, cautioned that while AI can enhance mathematical research, it may also lead to a loss of direct experience with mathematical understanding. Venkatesh emphasized the importance of preserving valuable aspects of mathematical culture as AI becomes more integrated into the discipline.
Furthermore, many mathematicians are transitioning from academia to work with technology firms like OpenAI and Google, or joining math-focused AI startups such as Harmonic and Axiom Math. Jeremy Avigad, director of the Institute for Computer-Aided Reasoning in Mathematics at Carnegie Mellon University, highlighted that the corporate interest in AI for mathematics stems from recognizing that combining machine learning insights with mathematical precision is key to achieving general intelligence.
Recent Milestones in AI and Mathematics
By early 2026, the initial shock at AI’s capabilities had evolved into a sense of wonder. A significant challenge called “First Proof” was launched in February, where participants had one week to use AI models to solve ten research-level questions across various mathematical fields. The selected questions were designed to be unlikely to have appeared in the algorithms’ training data. Remarkably, the models succeeded in solving over half of the problems, marking a significant milestone in AI’s journey through advanced mathematics.
In a blog post analyzing the results, Daniel Litt remarked, “It’s very likely that this technology is bigger than the computer.” This sentiment reflects a growing recognition of AI’s potential to revolutionize mathematics.
The Evolution of AI in Mathematics
The advancements seen in 2025 did not occur in isolation. Pushmeet Kohli, vice president of science at Google DeepMind, noted that efforts to solve math problems with AI began as early as 2018. François Charton, now at Axiom, started exploring machine learning applications in mathematics in 2019. Initially, these efforts focused on solving known problems to test the efficacy of new techniques. By 2024, researchers began to explore problems with rich datasets, using AI to construct mathematical objects with quantifiable properties.
In January 2025, Terence Tao and Javier Gómez-Serrano from Brown University collaborated with DeepMind mathematicians Adam Wagner and Bogdan Georgiev on a project called AlphaEvolve. This AI system utilized Gemini to write extensive Python programs, which were then refined using genetic algorithms to discover optimal solutions to mathematical problems. The team learned to improve their prompts for AlphaEvolve, finding that positive reinforcement enhanced the model’s performance. Gómez-Serrano noted, “It worked better when we were prompting with some positive reinforcement to the LLM. Like saying ‘You can do this’ — this seemed to help.”
Conclusion
The integration of AI into mathematics is reshaping the field in unprecedented ways. As mathematicians embrace these powerful tools, the future of mathematics promises to be more collaborative and innovative. While challenges and concerns remain, the potential for AI to enhance mathematical research is undeniable.
Note: This article reflects the ongoing developments in the intersection of artificial intelligence and mathematics as of April 2026.
