超越二元：开放权重高级AI的细致路径
Beyond the Binary: A nuanced path for open-weight advanced AI
AI总结：
本文提出了一种基于安全评估的分层模型发布方法，旨在解决开放权重高级AI模型在安全性和监管方面的挑战。
摘要：
开放权重高级AI模型——参数可自由下载和调整的系统——正在重塑全球人工智能格局。随着这些模型迅速缩小与封闭替代品的性能差距，它们使突破性研究成为可能，并扩大了强大工具的可及性。然而，一旦发布，它们便无法收回，其内置的安全措施可能通过微调或劫持绕过，这些风险当前的治理框架无法有效应对。本报告超越了“开放”与“封闭”二元框架。我们评估了开放权重高级AI的当前状况，检查了技术能力、风险概况以及欧盟、美国、中国、英国和国际论坛的监管回应。我们发现开发者和司法管辖区在安全实践上存在显著差异，没有普遍采用的标准来确定何时或如何公开发布高级模型。我们提出了一种分层、以安全为依托的方法来发布模型，其中开放性由严格的风险评估和已证明的安全性决定，而非意识形态或商业压力。我们为开发者、评估者、标准制定者和政策制定者提出了可操作的建议，以实现负责任的开放，同时投资于技术安全和社会准备。
英文摘要：
Open-weight advanced AI models -- systems whose parameters are freely available for download and adaptation -- are reshaping the global AI landscape. As these models rapidly close the performance gap with closed alternatives, they enable breakthrough research and broaden access to powerful tools. However, once released, they cannot be recalled, and their built-in safeguards can be bypassed through fine-tuning or jailbreaking, posing risks that current governance frameworks are not equipped to address. This report moves beyond the binary framing of ``open'' versus ``closed'' AI. We assess the current landscape of open-weight advanced AI, examining technical capabilities, risk profiles, and regulatory responses across the European Union, United States, China, the United Kingdom, and international forums. We find significant disparities in safety practices across developers and jurisdictions, with no commonly adopted standards for determining when or how advanced models should be released openly. We propose a tiered, safety-anchored approach to model release, where openness is determined by rigorous risk assessment and demonstrated safety rather than ideology or commercial pressure. We outline actionable recommendations for developers, evaluators, standard-setters, and policymakers to enable responsible openness while investing in technical safeguards and societal preparedness.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
