Core Insights - China has made significant advancements in AI research and deployment, particularly in generative AI technology, with predictions indicating a shift towards open GenAI models and a focus on cost-effective AI solutions by 2025 [1][2] - There is a growing concern regarding the return on investment in AI, with only 13% of respondents expressing strong confidence in the ROI, leading to a more cautious investment environment [1][2] Group 1 - Gartner predicts that by 2027, over half of Chief Data and Analytics Officers (CDAOs) will secure dedicated budgets for data literacy and AI literacy programs due to unmet expectations in GenAI investments [1] - The AI industry is experiencing a rationalization phase, with a shift in investment focus towards value and application, despite the overall heat of the industry remaining stable [2] - By 2028, demand for AI development skills in Chinese enterprises is expected to grow by 50%, with 33% of enterprise software incorporating agent-based AI by 2028, up from less than 1% in 2024 [2] Group 2 - The rise of Agent AI has improved productivity, but there remains a significant gap between market expectations and the current capabilities of general-purpose intelligent agents [3] - The current AI ecosystem is characterized by blurred boundaries, fostering an inclusive development model, particularly in China, where unique enterprise data is becoming crucial for differentiating AI application effectiveness [3]

360公司创始人周鸿祎表示，今年人工智能的发展，特别是大模型的发展，确实给人们带来了惊喜，但 目前还处于初级阶段。如果今年人工智能不能进化到智能体，那就是一场泡沫和闹剧。周鸿祎认为，今 年人工智能的发展，特别是大模型的发展，确实给人们带来了惊喜，但目前还处于初级阶段。周鸿祎认 为，今年人工智能的发展，特别是大模型的发展，确实给人们带来了惊喜，但目前还处于初级阶段。 (证券时报) ...

大厂押注志愿服务 2025.06. 25 本文字数：2999，阅读时长大约5分钟 作者 | 第一财经 陈杨园 高考成绩陆续揭晓，志愿填报成为考生与家庭面临的又一份高考"答卷"。 过往这个市场里的焦点有张雪峰、高考志愿填报师。记者在电商平台发现，花费300元至700元，消 费者就能够获得许多"高考志愿填报指导师"的认证，有的开课七天考试，半个月拿证，还有的甚至不 用上课，一两天就能出证，商家对记者表示，"没有条件，只要年满18周岁。有题库，考试包过。" 找到一个靠谱的志愿填报指导师不容易，高成本、高风险推动更多需求涌向互联网大厂的智能体。清 华大学研究团队近日发布的一项测评显示，研究团队从市场采买了30份价位在1000-5000元的志愿 专家报告，发现有AI产品与志愿填报专家水平相当。 AI在志愿填报中究竟能扮演什么角色？决策依据是什么？边界又在哪里？这是现在要用好AI工具之 前需要先回答的问题。 2025年，互联网大厂的AI志愿填报工具可谓百花齐放。夸克发布国内首个为高考志愿填报场景开发 的高考志愿大模型；百度上线"高考高频考点库""AI志愿助手"等系列产品；QQ浏览器上线行业首个 高考 Agent"AI高考 ...

金十数据6月25日讯，在夏季达沃斯论坛期间，360集团创始人周鸿祎表示，当前人工智能发展已经进入 下半场，智能体成为主角。"如果只把大模型当作工具来用，或许只能提升30%、50%的效率；但当大 模型进化为智能体，使其像数字助理一样帮人们处理各种复杂工作，人的角色就会转变为领导智能体、 规划人工智能、管理人工智能，人也将变为超级个体。因此近期的全球创业热潮包括中国的创业热潮， 也因为人工智能重新卷起了一个高潮。"周鸿祎成，"如果今年人工智能不能进化到智能体，那这次人工 智能可能又是一场泡沫，就不是工业革命了，而是一场闹剧。所以我们非常幸运地渡过了这一关。" 周鸿祎：当大模型进化为智能体 人也将变为超级个体 ...

6月19日，鼎捷数智股份有限公司于武汉举办2025(第四届)数智未来峰会，以"AI创变·破界生长"为主 题，汇集政产学研多方专家学者、生态伙伴、行业媒体、企业代表等近1000名嘉宾参会，共同探索数智 经济时代的前沿路径与未来机遇，解锁AI加速驱动业务场景与生产力创新，重塑核心竞争力的成功范 式。 鼎捷数智董事长兼总裁叶子祯邀请其数智分身共同完成本届峰会开场。数智分身通过作诗、学习方言等 精彩演绎，全力呈现AI与人的深度合作，诠释"科技让未来变得更加美好"的丰富畅想。 叶子祯提出"分身非替身"的主张，强调"AI的价值，不在于替代人，而在于让人从重复劳动事务中解 放，并为人创造全新的生产力"。鼎捷致力用"1个模型、2种内核、3类智能体"创建一个AI+的时代，实 现"智慧工作，乐享生活"。 于"九省通衢"之腹地 共绘AI创变之华章 "九省通衢连算力，千载商都拓云疆。若问创变何所向?鼎捷共绘赋能章。"叶子祯的数智分身赋诗一首 道出了本届峰会落子武汉的原因及峰会主题——武汉在地理位置上是内陆腹地最大的综合交通枢纽、国 家经济地理"心脏"，在中部地区崛起乃至构建新发展格局中占据至关重要的地位。在数智化发展布局 上，武汉在 ...

2025年被认为是智能体落地的元年，也是爆发年。无论是个人生活中的智能助手，如旅游规划、健身管家、文案生成智能体，亦或是企业端的数字员工，如 智能客服、供应链管理、金融投顾智能体，都在2025年不断涌现。 但繁荣背后，无论是消费级智能体或是企业级智能体都面临不少难题和挑战，其中之一就是内容安全问题。AI输出的文字、图片是否合规，会不会"乱说 话"，让用户体验大打折扣？ 今日，网易易盾内容安全插件上线国内各大智能体开发平台，包括Coze、Dify、千帆App Builder、文心智能体平台、腾讯云智能体开发平台及网易 CoreAgent智能协作中枢等，智能体开发者只需一键调用，就能在智能体内添加内容安全检测能力。这也是国内首个在智能体领域被大范围使用的AIGC内容 安全插件，将在未来持续赋能智能体生态健康发展。 4月以来，已有3500余款AI产品被处置 据IDC预测，Agent中国市场规模预计将突破300亿美元。Gartner数据则显示，到2028年，至少15%的日常工作决策将通过AI智能体自主做出，1/3的企业级软 件应用当中都会包含智能体。 智能体依托AI大模型生成多样化内容，其中可能包含虚假信息、敏感词汇 ...

Core Viewpoint - The application of large models in the financial industry is transitioning from an exploratory phase to a practical phase, becoming a necessity rather than an option [2][3]. Group 1: AI Integration in Financial Institutions - Financial institutions are increasingly integrating large models into their core business processes, moving beyond auxiliary tools [2]. - The current trend shows that AI applications in finance are shifting from customer service to core business areas such as wealth management and insurance claims [3]. - The year is being referred to as the "Agent Year," indicating a significant evolution in AI capabilities from digital assistants to digital employees [3]. Group 2: Challenges in AI Implementation - Financial institutions face challenges with large models, including a lack of understanding of financial contexts and concerns about data safety and compliance [3][4]. - There is a need for a specialized financial model rather than generic models, which are often seen as inadequate for the complexities of the financial sector [4]. Group 3: Successful AI Implementation Factors - Successful implementation of financial AI requires a specialized financial model, a responsive knowledge base, and the ability to facilitate business analysis and decision-making [4]. - Ensuring safety, compliance, and professionalism in financial models is crucial for creating effective financial intelligent agents [4]. Group 4: Pathways for AI Deployment - Ant Group has identified four pathways for AI deployment in financial institutions: building a model platform, creating AI-native mobile banking services, applying models in business scenarios, and prioritizing model deployment as a key project [5]. - The company offers flexible service models, including private deployment, SaaS subscriptions, and performance-based billing [5]. Group 5: Collaboration and Innovation - Ant Group plans to launch over a hundred intelligent agent solutions across various financial sectors, including wealth management and risk control [6]. - The integration of AI into business processes is seen as a strategic opportunity for financial institutions to drive organizational upgrades [6]. Group 6: Future of Financial AI - The development of financial AI is viewed as a long-term process requiring continuous iteration and improvement [11]. - Ant Group is working on creating independent financial models to bridge the gap between generic models and the specific needs of financial institutions [19]. Group 7: Data Security and Knowledge Management - Data security concerns are addressed through methods such as data anonymization and hybrid model deployment [17]. - The importance of a unified knowledge base is emphasized, as fragmented knowledge can hinder the effectiveness of AI applications in finance [18]. Group 8: Ecosystem Collaboration - Ant Group is merging its AI and cloud services to enhance product interoperability and address the challenges faced by financial institutions [20]. - The company aims to provide a comprehensive AI product system that considers both technical and business aspects of AI implementation [20].

机器之心报道 机器之心编辑部 恭喜获奖者。 RSS（Robotics: Science and Systems，机器人科学与系统会议）是机器人领域顶级学术会议，自 2005 年起每年举办一次，该会议旨在促进机器人领域的科学研究和 技术应用的发展。 地址：https://roboticsconference.org/program/awards/ 杰出 Demo 论文奖 论文标题：Demonstrating MuJoCo Playground 论文摘要： 该研究提出了 MuJoCo Playground—— 这是一个基于 MJX 构建的完全开源机器人学习框架，其核心设计目标是大幅简化仿真环境搭建、模型训练以及 仿真到现实场景的迁移全流程。研究人员仅需执行简单的「pip install playground」安装命令，即可在单 GPU 硬件上完成分钟级策略训练。 该框架支持四足机器人、人形机器人、灵巧手及机械臂等多类型机器人平台，能够直接基于状态观测或像素级输入实现零样本仿真到现实迁移。 今年大会已于 6 月 21 日至 25 日在美国洛杉矶举行。杰出 Demo 论文奖、杰出系统论文奖、杰出学生论文奖、杰出论文奖 ...

Group 1 - The core viewpoint of the articles highlights a shift in the AI industry from large model development to application-focused strategies, with companies like DeepSeek and Manus leading the way in this transition [1][5][7] - The investment logic in the AI sector has changed, with a focus on application investments rather than foundational model investments, as evidenced by the reduced financing amounts and the cautious approach of investors [6][7][9] - The "AI Six Tigers" have shown varied commercial progress, with companies like Zhipu and Zero One Wanwu making strides in B-end applications, while others like MiniMax and Moon Shadow focus more on C-end applications [9][10][11] Group 2 - DeepSeek has established itself as a dominant player, with significant backing and no immediate need for external financing, while other companies in the "AI Six Tigers" have struggled to secure new funding [6][8] - The emergence of new models from competitors like MiniMax and Moon Shadow indicates a competitive landscape where companies are striving to outperform DeepSeek [2][3] - The trend towards intelligent agents has become a consensus among AI companies, with multiple firms launching their own agent products in response to market demands [4][11] Group 3 - Companies are increasingly focusing on building differentiated competitive barriers in vertical markets to survive the ongoing industry reshuffle [1][12] - The commercial viability of AI applications is being tested, with a notable emphasis on B-end markets as a more sustainable path for revenue generation compared to C-end markets [11][12] - The overall investment landscape is evolving, with a greater emphasis on practical applications of AI technology across various industries, reflecting a broader market demand for AI solutions [7][12]

我们似乎距离这样的场景一望可及：在某一天的夜色中，一架无人巡检机在高空低鸣，摄像头精准锁定了主控泵房的机械抖动异常。与此同时，地面上的 四足机器人接收到异常码，避开障碍物迅速前往现场。两者并非通过云端调度，而是在本地通过"边缘智能体操作系统"自组织形成任务协同：无人机负责 视觉识别与路径分析，地面机器人完成执行与反馈。整个过程无需人的干预，也无需连接远程云平台。 这不是科幻，而是边缘AI从推理引擎迈向协作智能体的真实演进。 过去几年里，边缘AI的演进路径清晰可见——从最初的TinyML微型机器学习探索低功耗AI推理，到边缘推理框架的落地应用，再到平台级AI部署工具的 兴起，以及最近大热的垂类模型，我们已经实现了"让模型跑起来"的任务。 然而，边缘AI的下一步，不是继续堆叠更多模型、更多参数，而是回答一个更根本的问题：当AI模型跑起来之后，它们能否协作起来？ 这一局限，正是边缘AI走向更高智能形态的"隐形天花板"。 真正的边缘智能，不止于做出判断，而是要做出决策、组成系统、执行任务。这正是边缘AI从静态推理向动态智能体演进的起点。 我们需要的不再是一个更大的模型，而是一群能协作的模型。可以认为，模型让设备看见世 ...