Ghosh's intelligent infrastructure proposal envisions AI as coordinator for science

Satrajit S. Ghosh's paper outlines a vision for an intelligent infrastructure that transforms how scientific research is conducted, using neuroscience as a demanding test case. Traditional infrastructure is criticized as static, fragmented, and unsustainable—creating data silos, reproducibility issues, and short-lived solutions. The exponential growth of multimodal, multiscale data in neuroscience, paired with urgent clinical needs, demands an adaptive system that can expose computable states, coordinate across systems, and improve through use.

The author argues that AI's role must shift from assistive (analysis) to transformative (coordination). Ghosh calls for a decentralized, self-learning, and self-correcting ecosystem where humans and AI collaborate seamlessly. The paper builds on existing principles for data, collective benefit, and digital repositories, offering operational guidelines to implement these. Key challenges include chronic underfunding of research infrastructure, the need to acknowledge diverse contributions beyond publications, and global coordination. If realized, this blueprint could accelerate discovery, ensure reproducibility and ethical practices, and set a precedent for other scientific domains.