AI's Mixed Impact on Mathematics Raises Hopes and Fears, Report Says

The report quotes The Verge's AI reporter as saying that AI "went from being very terrible to seemingly genuinely quite good at a professional level in a very short space of time." The reporter noted that academic math papers often do not contain numbers, allowing AI to excel at abstract reasoning despite its glaring gaps in arithmetic.

At a certain critical mass, the reporter explained, AI systems become capable of forging connections between different areas and applying old methods in new ways, similar to what has been observed in writing and programming. Newer models appear to have reached that level, producing work that is on par with good mathematicians in some domains while remaining hopeless at basic counting.

The Slashdot report outlines concerns among researchers that AI could fundamentally alter employment structures and leave the discipline "quite sterile." Some mathematicians fear that the breakthroughs leading to surprising new directions may cease, and graduate students have posted bleak essays online questioning their future role. One worry is that researchers could become "glorified AI proof checkers," a prospect many find unsatisfying.

On the question of democratizing access to high-level mathematical proofs, the report says many mathematicians are weary. They are "fed up" with AI-generated or AI-assisted papers flooding publications and preprint servers, with some receiving multiple emails each week from people who believe they have solved major problems using tools like ChatGPT or Claude but who lack the mathematical skills to verify their work.

Yet the report also highlights positive examples. One mathematician noted that a talented undergraduate, aided by AI, produced a legitimate paper at the graduate level. Others suggested that global access to these tools could broaden participation in advanced mathematics, although the high cost of running such systems remains a significant barrier.