China Daily reports AI accelerates quantum computing, experts now predict.

For four decades, error correction was a fundamental obstacle to building practical quantum computers, but artificial intelligence is now offering promising solutions. Andrew Chi-Chih Yao, Turing Award winner and Dean for Interdisciplinary Information Sciences at Tsinghua University, identifies AI for science as “the most interesting, important and promising direction for AI research over the next three to five years.” Yao notes that AI and quantum technologies are expanding the frontiers of human knowledge, with quantum AI emerging and the potential for significant progress over the next five to 10 years. Su Hao, inaugural dean of the Institute of General Physical Intelligence at Fudan University, proposes a solution to the problem of hallucinations in large language models: grounding them in physical reality by making predictions, taking action, and learning from reality’s feedback.

Yao, also an academician with the Chinese Academy of Sciences, highlights the potential for AI to accelerate progress in areas previously considered intractable, specifically citing the emergence of quantum AI as a frontier beyond present-day capabilities. The convergence of AI and quantum technologies expands the boundaries of knowledge, and is crucial because AI algorithms, while powerful, are fundamentally constrained by the laws of physics and mathematics; therefore, breakthroughs require tools that can operate within, and even extend, those boundaries. The application of AI isn’t limited to overcoming technical hurdles, but also offers a new approach to addressing fundamental limitations within AI itself.

A shift in focus is underway within artificial intelligence research, moving beyond purely computational advancements toward systems grounded in physical interaction and demonstrable reliability. This emphasis stems from the recognition that even sophisticated algorithms are ultimately constrained by the fundamental laws governing the physical world. According to Hao, the industry will increasingly prioritize reliable system operation over spectacular demonstrations, with generalizability as the ultimate goal and reliability as the essential starting point. This signals a move away from simply scaling model size and toward building systems that consistently perform as expected in real-world scenarios, demanding a focus on knowledge aggregation rather than architectural innovation.