Rand兰德：2025年保障前沿人工智能（AI）的治理路径研究报告（英文版）

Growing concerns about the societal risks posed by advanced artificial intelligence (AI) systems have prompted debate over whether and how the U.S. government should promote stronger security practices among private-sector developers. Although some companies have made voluntary security commitments, competitive pressures and inconsistent approaches raise questions about the adequacy of self-regulation. At the same time, government intervention carries risks: Overly stringent security requirem