Science
Physics Model Reveals Limits of AI Energy Efficiency
An international research team led by physicist Hans Briegel at the University of Innsbruck has uncovered a fundamental physical trade-off between prediction accuracy and energy consumption in intelligent systems. Published in Physical Review X, the study examines the physical limits governing how artificial intelligence and other cognitive systems process information. As AI becomes more deeply integrated into daily life through digital assistants and advanced robotics, understanding its thermodynamic constraints is increasingly critical. The researchers applied principles of theoretical physics to map out how computational processes inherently require energy to make accurate environmental predictions.