Will AI progress substantially displace the role of mathematicians?

Determine whether continued progress in artificial intelligence systems for mathematical research, exemplified by models such as GPT-5, will eventually substantially displace the role of human mathematicians in conducting mathematical research.

Background

The paper reports a controlled experiment using GPT-5 within the Malliavin–Stein framework, leading to quantitative results in both Gaussian and Poisson settings. Beyond the technical contributions, the authors reflect on the broader implications of rapidly advancing AI systems for mathematical practice and research workflows.

Based on their experience—where the AI acted effectively as an executor under close human guidance but made critical mistakes without oversight—the authors raise a broader, unresolved question about the long-term impact of such progress on the profession and role of mathematicians.