留守儿童
OpenAI奥特曼将向美国官员介绍下一代AI模型，美国AI审查框架预计几周内完成_我的网站
A | OpenAI首席执行官奥特曼计划于下周赴华盛顿，就即将发布的新一代人工智能模型向特朗普政府官员及国会议员进行报告。此次访问的时机，正值美国加快推进AI安全审查机制建设。 Beijing, Oct 12 (UNI) The Chinese military is allegedly benchmarking the practices used in current hostilities in Europe to develop a strategy of "hybrid war" against Taiwan, Chen Ming-tong, Director-General of the Taiwanese National Security Bureau, said on Wednesday.
"This year, the army of Communist China, borrowing the experience of military operation in Europe, started to develop a hybrid war against Taiwan and boosted its military training," Chen told the local parliament.
The Taiwanese official also said that China continued to send its aircraft and vessels to the Taiwanese air defense identification zone and carried out "other provocations" after conducting large-scale military exercises in early August.
Chen added that Beijing's "threatening" actions, which include the attempts to change status quo in the Taiwan Strait, "jeopardize peace and stability" in the Asia-Pacific region.
In late August, the Taiwanese military fired on a Chinese People's Liberation Army (PLA) drone that approached the Kinmen Islands. This was the first time that an actual shot was fired instead of warnings.
The situation around Taiwan escalated after US House Speaker Nancy Pelosi visited the island in early August. Beijing condemned Pelosi's trip, which it regarded as a gesture of support for separatism, and launched large-scale military exercises in the vicinity of the island. Despite this fact, several countries, including France, the United States, Japan and others, have sent their delegations to the island since then, further increasing tensions in the Taiwan Strait.
Taiwan has been governed independently from mainland China since 1949. Beijing views the island as its province, while Taiwan — a territory with its own elected government — maintains that it is an autonomous country but stops short of declaring independence. Beijing opposes any official contacts of foreign states with Taipei and considers Chinese sovereignty over the island indisputable.
UNI/SPUTNIK GNK。
OpenAI全球公共事务负责人Chris Lehane周二表示，奥特曼此次访问之际，美国官员正在制定一套AI安全审查框架，预计将在未来几周内完成。
他表示，奥特曼此次会晤的一部分内容将聚焦OpenAI新的模型系列，包括这些模型将如何影响就业和工作方式。
“我们认为，这一模型系列将带来一些非常有趣的能力，尤其是在工作场景以及扩大工作规模方面。”
Lehane的上述表态，距Kimi K3模型发布仅数日之隔。该模型被认为可与OpenAI、Anthropic等顶级AI产品相抗衡，引发市场对美国AI领先地位持续性的质疑，并导致科技股出现抛售。
B |
监管框架加速成形，独立审查机构方案浮出水面
美国AI监管体系的构建正在提速。特朗普政府正在评估设立一家独立监管机构的可行性，该机构将在产业界参与下对AI模型进行审核。财政部长贝森特参与制定的方案已被纳入考量，旨在为AI企业提供更明确的合规预期。
上述方案在方向上与谷歌DeepMind首席执行官Demis Hassabis上周提出的构想相近。
C | Hassabis本周亦赴华盛顿与政策制定者会面，就其监管愿景展开游说。
D |
当前监管讨论的背景，在于硅谷对近期美国政府临时性、碎片化管控措施的不满持续升温。上月，Anthropic受到美国出口管制冲击，被迫暂时下线其Fable 5和Mythos 5模型。
OpenAI虽未受到类似的公开管制措施，但据悉在政府要求下对最新GPT-5.6模型作出了重大调整方可发布。
国会立法受阻，OpenAI转寻州级路径
在联邦层面推进统一立法的努力面临不确定性，OpenAI已着手寻求备选路径。
E | Lehane表示，若国会无法建立全国性AI治理标准，该公司将推动各州相互对标，以"逆向联邦主义"的方式形成实际上的统一规范。
F |
"如果无法推动国会建立这些全国标准，"他说，"另一条路就是我们所说的逆向联邦主义——推动不同州相互镜像对齐。"
目前，OpenAI正支持马萨诸塞州的一项立法提案，该提案要求主要AI开发商采取更多措施识别灾难性风险，Anthropic亦对该提案表示支持。
G |
Current article:http://6v1emyg.laonenpeidiebenhuaiyunnangshi.cfd/news/20260826_550994.html
Published on:09:31:05
