Key Points
- Physicists used AI as a tool for discovery, not just for processing data.
- The AI model learned from a small amount of data to find new physical laws governing dusty plasma.
- It precisely described “non-reciprocal” forces between particles, where attraction and repulsion aren’t mutual.
- The findings corrected older, inaccurate theories about particle charge and force.
In a major leap for scientific discovery, physicists have used artificial intelligence to uncover new physical laws, moving beyond AI’s typical role as a data-processing tool. In a new study, published in Proceedings of the National Academy of Sciences, an Emory University team used AI to reveal surprising new details about the forces at play in a “dusty plasma”—an ionized gas filled with tiny suspended particles, common in space and planetary rings.
The team designed a special neural network that could learn from a small amount of experimental data. Justin Burton, one of the lead researchers, explained that their AI is not a “black box” because they understand exactly how it works. He added that the framework is universal and could be used to study other complex systems.
The AI described the physics of dusty plasma with over 99% accuracy. It precisely modeled “non-reciprocal” forces between particles. Imagine two boats on a lake: the AI found that the wake from a leading particle attracts a trailing particle, but the trailing particle’s wake always repels the leading one.
The model also corrected long-held theories, showing that a particle’s electrical charge isn’t perfectly proportional to its size. This detail depends on the plasma’s density and temperature.
The researchers believe this AI approach can be a starting point for understanding a wide range of “many-body systems,” from industrial materials like paint to the collective movement of cancer cells in the human body. The goal is to use this tool to discover the fundamental rules that govern how complex groups of particles interact and move together.
