🇨🇳 China AI News Daily — 2026年03月18日
🇨🇳 China AI News Daily
2026年03月18日 · 15 articles
🤖 New AI Model & Product Releases
MiniMax发布新一代大模型M2.7 [MiniMax Releases Next-Generation Model M2.7]
钛媒体 · 2026-03-18
(广角观察)【MiniMax发布新一代大模型M2.7】钛媒体App 3月18日消息,MiniMax发布新一代Agent旗舰大模型M2.7,首次展示'模... [MiniMax has released its next-generation flagship agent model M2.7, marking its first public demonstration of new model capabilities.]
智谱发布龙虎专用模型,大模型进入场景定制阶段 [Zhipu Releases Specialized Model GLM-5-Turbo, Signaling Era of Scenario-Customized AI]
i黑马 · 2026-03-18
智谱发布了一款名为GLM-5-Turbo的新模型.这个模型很特别,它... 连续涨价的姿态,在一年多来习惯了'价格战'的大模型市场中显... [Zhipu AI has released GLM-5-Turbo, a specialized scenario-specific model, signaling a shift from the industry's prolonged price war era toward customized deployments.]
智己超级智能体正式发布,基于阿里千问大模型 [IM Motors Launches 'IM Ultra Agent' Super AI System Powered by Alibaba Qwen]
问驾 · 2026-03-18
点击上方蓝字关注我们吧~今晚,智己汽车发布了'IM Ultra Agent'超级智能体,集成千问大模型,实现线控底盘、L4级辅助驾驶大模... [IM Motors has officially launched its 'IM Ultra Agent' AI system integrating Alibaba's Qwen large model, enabling drive-by-wire chassis control and L4-level assisted driving.]
日本最强'国产大模型',被扒全是中国代码 [Japan's 'Strongest Domestic AI Model' Rakuten AI 3.0 Found to Be Based on Chinese DeepSeek V3 Code]
观察者网 · 2026-03-18
'最强日本国产大模型'根据乐天的官方新闻稿,Rakuten AI 3.0是"基于开源社区最优秀的模型,结合乐天独自的高品质双语数据、... [Rakuten AI 3.0, marketed as Japan's strongest domestic large model, has been revealed to be nearly identical to China's DeepSeek V3 based on matching architecture configurations.]
国产大模型周调用量再超美国 [Chinese Domestic AI Models Surpass US Models in Weekly API Call Volume for Second Consecutive Week]
第一财经 · 2026-03-18
国产大模型周调用量已连续两周超过美国大模型产品.数据显示,3月9日当周,全球模型调用量排名前九名中,公开明确的国产模型占... [Chinese domestic large language models have surpassed US counterparts in weekly API call volumes for two consecutive weeks, with domestic models dominating the global top-nine rankings as of the week of March 9.]
王兴出招,美团上线首款AI产品 [Wang Xing Makes His Move: Meituan Launches Its First AI Product 'Wow']
互联网头条 · 2026-03-18
在五个月后推出了首个AI产品.Tech星球获悉,美团近期上线了一款名为'Wow'的独立APP,已完成产品备案,这是美团首个AI交互... [Meituan has launched its first standalone AI product, an interactive app called 'Wow', which has completed China's mandatory product registration process.]
盘古大模型全新版本即将发布,建议关注华为产业链标的 [Huawei's Pangu Model New Version Imminent; Analysts Recommend Watching Huawei Supply Chain Stocks]
通信万花筒 · 2026-03-18
核心观点华为云盘古大模型全新版本即将发布.目前,华为云已推出五大水平领域的基础大模型(L0),分别为:NLP(中文语言)大... [A new version of Huawei Cloud's Pangu large model is imminent, with analysts flagging it as a catalyst for Huawei's broader supply chain ecosystem.]
百度发布首个量子领域大模型,5分钟内可完成上万字专利文档 [Baidu Launches First Quantum-Domain Large Model, Capable of Generating 10,000-Character Patent Documents in 5 Minutes]
百度 · 2026-03-18
加速量子技术与大模型深度融合.他还发布了量子领域大模型白皮书,展望量子领域大模型的未来发展趋势和技术潜力.当前,量子计... [Baidu has released its first large model specialized for the quantum computing domain, along with two native AI applications and a white paper outlining the technology's future potential.]
📜 AI & Technology Governance Policies
政策解读|工信部'人工智能+'政策:以'人工智能+制造'为核心抓手,打造新质生产力、赋能新型工业化 [Policy Interpretation | MIIT's 'AI+' Policy: Using 'AI+Manufacturing' as Core Lever to Build New Productive Forces and Empower New Industrialization]
仙人论数 · 2026-03-18
以'人工智能+制造'为核心抓手,配套算力、数据、网络、安全等全链条政策,打造新质生产力、赋能新型工业化.一、顶层框架与... [An analysis of China's Ministry of Industry and Information Technology 'AI+' policy framework, centering on AI-driven manufacturing with supporting policies across computing power, data, networking, and security to develop new productive forces.]
生成式人工智能服务管理暂行办法公布 [Interim Measures for the Management of Generative AI Services Published]
澎湃新闻 · 2026-03-18
孙铭蔚近日,国家网信办联合国家发展改革委、教育部、科技部、工业和信息化部、公安部、广电总局公布《生成式人工智能服务管理... [China's Cyberspace Administration, jointly with six other ministries, has officially published the Interim Measures for the Management of Generative AI Services, establishing a regulatory framework for generative AI.]
李强:中国为全球人工智能治理提供建设性方案——《全球人工智能治理倡议》解读 [Li Qiang: China Provides Constructive Solutions for Global AI Governance — Interpreting the 'Global AI Governance Initiative']
中国网信杂志 · 2026-03-18
此外,在人工智能治理领域,中国外交部已先后面向国际社会发布《中国关于规范人工智能军事应用的立场文件》《中国关于加强人工... [An analysis of China's Global AI Governance Initiative, highlighting China's diplomatic positions on regulating military AI applications and strengthening international AI cooperation frameworks.]
人工智能面对多重安全风险,协同治理机制有待健全 [AI Faces Multiple Security Risks; Collaborative Governance Mechanisms Need Strengthening]
第一财经 · 2026-03-18
加快形成动态敏捷、多元协同的人工智能治理格局.同年8月,工业和信息化部等部门联合发布《人工智能科技伦理管理服务办法(试... [China's First Financial Daily examines multiple security risks facing AI and calls for strengthening multi-stakeholder collaborative governance mechanisms, noting recent ministry-level ethics management regulations.]
傅莹:人工智能与国际安全治理路径探讨 [Fu Ying: Exploring Pathways for AI and International Security Governance]
人民论坛网 · 2026-03-18
其他国家也表示出对人工智能应用的安全担忧,说明人工智能治理是人类共同的难题,不是一国两国能够解决的.中美开展对话与合作... [Former Chinese diplomat Fu Ying argues that AI governance is a shared global challenge requiring multilateral cooperation, including US-China dialogue, to address international security risks.]
静水流深:美国人工智能治理的特征、趋势与启示 [Still Waters Run Deep: Characteristics, Trends, and Lessons of US AI Governance]
阿里研究院 · 2026-03-18
二、美人工智能治理呈融合型、渐进性特征,行政治理为先,软法约束为主,对刚性立法形成了塑造和牵制效应(一)治理内涵脱胎于... [Alibaba Research Institute analyzes US AI governance as incremental and hybrid in nature, led by administrative measures and soft-law constraints that shape and restrain harder legislation.]
算法监管:如何进行算法备案?不备案的后果是什么?|星川ASTROLEGAL [Algorithm Regulation: How to File Algorithm Registration? What Are the Consequences of Non-Filing?]
星川ASTROLEGAL · 2026-03-18
除了自上而下的监管之外,还需自下而上、由内到外的自律与引领,平台作为算法服务的提供者,在通过算法获利的同时,还需意识到... [A legal explainer on China's algorithm registration (备案) requirements, covering the filing process, platform responsibilities, and penalties for non-compliance.]
Sources: WeChat public accounts · 36氪 · 量子位 · FT中文网 · 纽约时报中文网
Don't miss what's next. Subscribe to China AI News Digest: