In a preprint posted in November 2025, John P. Nelson, Olajide Olugbade, Philip Shapira, and Justin B. Biddle examine researchers’ perspectives on the role of artificial intelligence (AI) in manufacturing and materials science.
The study, “Can Artificial Intelligence Accelerate Technological Progress? Researchers’ Perspectives on AI in Manufacturing and Materials Science,” addresses the gap between widespread expectations that AI will dramatically accelerate innovation and the limited empirical evidence about how AI is actually used in research practice.
Drawing on 32 interviews with U.S.-based academic researchers experienced in AI and machine learning (ML), the authors find that AI is primarily employed to model materials and manufacturing processes and to enable faster, less costly exploration of design spaces. Interviewees report meaningful gains in efficiency, including reductions in development time, computational demands, and research costs.
At the same time, researchers emphasize important limitations. AI and ML tools tend to perform reliably only within well-characterized design spaces supported by dense datasets. Effective use requires significant expertise and careful integration with established empirical and theoretical methods. Some interviewees also express concern that reliance on AI-driven approaches may bypass opportunities for disruptive theoretical breakthroughs.
Overall, the findings suggest cautious optimism. AI and ML appear well suited to accelerating incremental, sustaining innovations in manufacturing and materials science. However, continued support for conventional empirical, computational, and theoretical research remains essential to sustain the conditions necessary for transformative advances.
The preprint is available on arXiv at https://arxiv.org/abs/2511.14007
Philip Shapira 