Artificial Intelligence in Research (tmodrzyk.github.io)

In a reflective journey through the evolution of artificial intelligence, a researcher highlights the dramatic changes in AI's capabilities, particularly in solving complex mathematical problems. As they complete their PhD thesis in September 2026, they recount how AI models, previously seen as limited in intelligence, have now successfully tackled the Navier–Stokes problem, a longstanding Millennium Prize challenge. The researcher notes a significant shift in their own work—moving from intensive mathematical problem-solving and coding to supervising AI agents that can now autonomously execute tasks, including drafting proofs and writing code in a fraction of the time it took them. This transformation signifies a pivotal moment for the AI/ML community, raising concerns about research quality amid an overwhelming surge in paper submissions and a shift from thorough evaluation to quantity-driven metrics. The post discusses how the proliferation of AI tools is leading to a decline in the rigor of peer review and a flood of low-stakes publications, undermining traditional measures of academic success. As researchers adapt to these changing dynamics, there is a growing need for an effective response to maintain the integrity of scientific inquiry, possibly through the adoption of AI-assisted review systems to manage the overwhelming volume of submissions.