Emory Physicists Use AI to Discover New Laws of Nature

ByMason Reed

April 28, 2026

Researchers at Emory University utilized a custom neural network to uncover hidden physical interactions within dusty plasma, overturning long-held assumptions about how particles behave in the universe’s most common state of matter.

In a significant leap for scientific inquiry, physicists at Emory University have demonstrated that artificial intelligence can do more than just process data—it can uncover fundamental laws of nature that have long eluded human observation. By training a specialized neural network on the behavior of ‘dusty plasma,’ the team has revealed new insights into the fourth state of matter, which comprises nearly 99.9% of the visible universe.

The research, published in the Proceedings of the National Academy of Sciences (PNAS), focused on the chaotic interactions of charged particles suspended in ionized gas. This environment, known as dusty plasma, is found in everything from the rings of Saturn to the smoke of terrestrial wildfires. Because these particles are governed by ‘non-reciprocal’ forces—where one particle might attract another while being simultaneously repelled by it—modeling their behavior has historically been a mathematical nightmare.

To solve this, first author Wentao Yu and senior co-authors Justin Burton and Ilya Nemenman spent over a year designing a ‘physics-tailored’ machine learning model. Unlike standard AI, which often operates as a ‘black box’ with no transparency, this model was built to respect the basic rules of physics while remaining flexible enough to discover unknown variables. The team used tomographic laser imaging in a vacuum chamber to track particle trajectories in 3D, providing the AI with high-fidelity experimental data.

The results were immediate and profound. The AI reached a 99% accuracy rate in describing particle interactions and, in doing so, overturned several established scientific assumptions. For decades, physicists assumed a particle’s electric charge was strictly proportional to its radius. The AI proved otherwise, showing the relationship is far more complex and dependent on local plasma density and temperature. It also corrected theories regarding how quickly forces drop off between particles of different sizes.

This breakthrough, which earned the 2025 Cozzarelli Prize from the National Academy of Sciences, represents a shift toward decentralized, high-tech innovation that empowers individual researchers. Supported by the National Science Foundation and the Simons Foundation, the methodology is designed to be universal. The researchers believe this framework can be applied to other complex systems, such as the movement of human cells or the behavior of industrial materials like ink and paint.

As the Trump administration recently overhauled the National Science Board to refocus American scientific priorities, projects like the Emory study highlight the value of principled, transparent research. By ensuring that AI tools are grounded in physical reality rather than bureaucratic algorithms, scientists are opening new frontiers in our understanding of the universe while maintaining the human oversight necessary for ethical discovery.

Leave a Reply

Your email address will not be published. Required fields are marked *