AI Can Propose an Answer. How Do We Ask Nature If It Is Right?
AI can write a program in seconds. With suitable tests, it can check the result, revise its approach, and try again. In mathematics, a formal proof can be checked by a computer. Fast, reliable digital verification helps explain AI's extraordinary progress in these fields.
But what happens when the answer is a new medicine, a safer battery, or a material with unfamiliar properties? A convincing explanation is only a beginning. Discovery must confront two fundamental limits.
The Physical Limit. Testing a scientific claim often requires contact with the physical or biological world. Samples must be prepared, reactions must proceed, and living systems need time to respond. AI may propose a thousand possibilities while a laboratory can test only a few. Robots and simulators can accelerate parts of this work, but simulations inherit assumptions and gaps in knowledge. Nature must still be consulted.The Quantum Limit. Much of nature operates according to quantum physics, while today's AI relies mainly on classical computers. A complete classical description of a general quantum state grows exponentially as the system grows. Useful approximations can take us far, but some problems demand a different approach. Quantum devices offer ways to simulate quantum behavior and, for certain tasks, learn directly from quantum information before it becomes an ordinary measurement record.In AI for Science and Quantum: The Ultimate Quest, Peter Y. Lee and and Yiyong Huang explore how classical AI, physical experimentation, and quantum computing might work together to overcome these obstacles. Through accessible explanations and examples-from protein prediction and robotic chemistry to simulated worlds and quantum experiments-the book follows the loop connecting questions, predictions, tests, and new understanding. It asks how AI's growing reasoning abilities can contribute to science, and what evidence is needed to move from recognizing patterns to understanding causes.
Written for curious readers without specialist training, the book examines both demonstrated achievements and the work still ahead. People remain responsible for the questions pursued, the authority delegated to machines, and the uses of discovery. Greater scientific power calls for greater integrity, compassion, and care in sharing its benefits.
The ultimate quest is to connect the power to imagine possibilities with the ability to learn from nature-and the wisdom to use what we learn.