英国获准测试最新AI模型的安全性，部长称 由威廉·詹姆斯 LONDON, 10月12日 - 英国人工智能安全研究所已获准使用最新的先进人工智能模型，并正与美国政府及主要开发商（包括谷歌）合作，评估其安全性，人工智能部长卡尼什卡·纳拉扬周一表示。
UK given access to safety-test latest AI models, minister says By William James LONDON, Oct 12 - Britain's AI Security Institute has been given access to the latest advanced AI models and is working with the US government and major developers, including Google, to assess their safety, AI Minister Kanishka Narayan said on Monday.
英国虽然没有任何主要人工智能开发商，但正寻求在评估和管理日益强大的人工智能系统所带来的风险方面发挥主导作用。
Britain, which is not home to any major AI developers, is seeking a leading role in efforts to assess and manage the risks posed by increasingly powerful AI systems.
自Anthropic和OpenAI等开发商披露人工智能代理出现失控或不受欢迎行为的案例以来，对人工智能安全的审查力度不断加强。一些业界人士警告称，该技术可能对人类构成生存威胁。
Scrutiny of AI safety has intensified after developers including Anthropic and OpenAI disclosed instances of rogue or undesired behaviour by AI agents. Some industry figures have warned the technology could pose an existential threat to humanity.
周一的声明未具体说明研究所正在评估的所有模型，但紧随上月有关白宫要求OpenAI和Anthropic暂停向英国测试者提供新模型的报道，直至完成美国方面的审查。
Monday's announcement, which did not specify all the models the institute is evaluating, comes after reports last month that the White House has asked OpenAI and Anthropic to withhold new models from British testers pending a US review.
“我可以确认，人工智能安全研究所的人工智能科学家目前正在对最新一代模型（包括Gemini 4 Argon）进行严格评估，以便及早发现风险、测试可靠的防护措施，帮助确保人们的安全，”纳拉扬表示。
"I can confirm that AI scientists at the AI Security Institute are now conducting rigorous evaluations of the latest generation of models, including Gemini 4 Argon, to identify risks early, to test robust safeguards, to help keep people safe," Narayan said.
在与主要人工智能开发商的自愿协议框架下，研究所可在模型部署前获得访问权限，从而评估其能力和风险，并为新兴的国际安全标准作出贡献。
Under voluntary agreements with major AI developers, the institute receives access to models before deployment, allowing it to assess their capabilities and risks and contribute to emerging international safety standards.
微软总裁布拉德·史密斯在BBC周一发表的采访中表示，英国在协助制定人工智能安全指南方面具有有利条件。
Microsoft President Brad Smith said in an interview published by the BBC on Monday that Britain was well placed to help lead work on AI safety guidelines.
9月发表的一封信中，AISI表示Anthropic已明确表明，除美国以外的任何组织均无法访问其Mythos 5.1模型。
In a letter published in September, the AISI said Anthropic had made clear that no organisation outside the US had access to its Mythos 5.1 model.
当被问及是否能够访问前沿人工智能系统——这是部长们此前从未确认研究所在更广泛范围内失去的情况——纳拉扬表示，政府与美国建立了基于信任的长期关系。
Asked about access to frontier AI systems — something ministers have never confirmed the institute had lost on a wider basis — Narayan said the government had a long-standing relationship with the US built on trust.
“当你依赖他人的技术时，必须清醒地认识到任何结果的永久性。但我非常有信心，由于我们进行的对话，我们将能够测试最新一代的模型。”他说道。
"When you're reliant on someone else's technology, you have to be clear-eyed about the permanence of any outcome. But I'm very confident that, as a consequence of the conversations we've had, we will be able to test the latest generation of models," he said.