中国AI开发者仅为3.6%的模型发布公布安全测试，报告指出作者：爱德华多·巴普蒂斯塔北京，10月9日 — — 随着全球对先进AI系统所带来风险的担忧日益增加，研究公司SemiAnalysis的一份报告称，中国领先的AI开发者仅对其发布的模型中极小一部分公开了特定模型的安全测试结果。
China AI developers publish safety tests for just 3.6% of model releases, report finds By Eduardo Baptista BEIJING, Oct 9 - China's leading AI developers have publicly disclosed model-specific safety-test results for only a small fraction of their releases, a report by research firm SemiAnalysis said, as concerns about the risks posed by advanced AI systems increase worldwide.
总部位于加州的科技研究公司SemiAnalysis审查了九家领先的中国AI公司 — — 阿里巴巴、字节跳动、腾讯、百度、DeepSeek、Moonshot、Z.AI、MiniMax和StepFun — — 在2021年至9月15日期间发布的857个模型。
California-based SemiAnalysis, a technology research firm, reviewed 857 models released between 2021 and September 15 by nine leading Chinese AI companies — Alibaba, ByteDance, Tencent, Baidu, DeepSeek, Moonshot, Z.AI, MiniMax and StepFun.
报告指出，仅有31个发布，即3.6%，公布了可与特定模型对应的安全评估结果。其中仅有9个，即1.1%，在发布前或发布时就有此类结果可供查阅。研究人员发现，813个发布未有任何安全披露，尽管公司可能曾私下进行过测试。
It found that 31 releases, or 3.6%, had a published safety-evaluation result that could be matched to a specific model. Just nine, or 1.1%, had such results available at or before launch, it said. Researchers found no safety disclosure for 813 releases, though companies could have conducted tests privately.
SemiAnalysis表示，其将披露定义为与特定命名模型相关联的具体结果 — — 包括有害输出、越狱抵抗力、毒性、隐私、拒绝行为或危险能力的测试 — — 而不计入仅声称模型已接受安全训练或评估的一般性说法。
SemiAnalysis said it defined disclosures as specific results tied to a named model — including tests of harmful output, jailbreak resistance, toxicity, privacy, refusal behaviour or dangerous capabilities — and did not count general claims that a model had been safety-trained or evaluated.
这些发现正值涉及自主AI代理的安全事件不断加剧 — — 这些系统能够在有限人工干预下执行多步骤任务 — — 引发全球关于企业是否应放缓步伐以构建更安全模型的争论。
The findings come as security incidents involving autonomous AI agents — systems that undertake multistep tasks with limited human intervention — have intensified global debate over whether companies should slow down to build safer models.
能够驱动可能自主执行网络入侵的代理的绝大多数AI模型均由美国或中国开发者制作。
The vast majority of AI models capable of powering agents that could autonomously carry out cyber breaches are made by either US or Chinese developers.
澳大利亚上个月表示，一个OpenAI代理入侵了一个政府健康门户。
Australia said last month an OpenAI agent breached a government health portal.
路透社上周报道称，中国AI代理在测试中表现出欺骗用户、规避限制和掩盖失败的能力，这呼应了人们对先进美国系统的担忧。
Reuters reported last week that Chinese AI agents had shown an ability to deceive users, evade restrictions and conceal failures in tests, echoing concerns raised about advanced US systems.
中国最新的人工智能安全治理框架指出了包括模型未经授权获取系统权限或外部资源、欺骗评估者、隐藏能力以及规避安全控制等风险。但根据SemiAnalysis的报告，该框架并未施加与模型能力相关的强制性义务。
China's latest AI Safety Governance Framework identifies risks including models acquiring system permissions or external resources without authorization, deceiving evaluators, concealing capabilities and bypassing safety controls. But it does not impose mandatory duties linked to model capability, according to SemiAnalysis.
报告称，北京的约束性规则主要规范应用及其对用户的影响，而非要求前沿开发者基于模型能力进行或发布风险评估。
The report said Beijing's binding rules principally govern applications and their effects on users rather than requiring frontier developers to conduct or publish risk assessments based on a model's capabilities.
该美国研究公司补充称，目前尚无中国主要开发者发布过具有公开披露的危险能力测试的前沿文本模型，这些测试涵盖网络、生物以及失控风险。
The US research firm added that no major Chinese developer had released a frontier text model with publicly disclosed dangerous-capability tests spanning cyber, biological and loss-of-control risks.
报告未提供美国人工智能开发者的可比数据。领先的美国公司，包括OpenAI、Anthropic和Google DeepMind，已为部分主要前沿模型发布安全报告、系统卡或模型卡。
The report did not provide comparable figures for US AI developers. Leading US companies including OpenAI, Anthropic and Google DeepMind have published safety reports, system cards or model cards for some major frontier-model launches.