本报记者 梁傲男 9月25日，主题为"EnjoyAI"的2025京东全球科技探索者大会在北京举行。京东集团股份有限公司（以下 简称"京东"）宣布，未来将持续加码技术投入，带动各个产业，形成万亿元规模的人工智能生态。 当前，人工智能正深刻重塑各行业供应链格局。依托在供应链及多产业场景中积累的技术与实践，京东 已在智算基础设施、模型构建、平台支撑、场景应用及产品落地等方面形成完整能力体系。 此次大会上，京东首次系统展示了京东的AI全景布局，不仅升级"JoyAI"大模型品牌，还发布了三款面 向消费者的产品、四大场景应用以及三大深度平台，致力于将自身的供应链核心能力更直接、更智能地 赋能至终端用户。 为产业创造价值 财报显示，今年第二季度，京东的研发开支达53亿元，同比增长25.7%。自2017年宣布全面向技术转型 以来，截至今年二季度末，京东体系的研发投入已累计超1500亿元。截至第二季度，京东供应链基础设 施资产规模近1700亿元，为技术应用提供了广阔的场景。 重塑供应链 在本次大会上，京东全面展示了其AI全景布局。目前，京东已在智算基础设施、模型、平台、场景及 产品等多个层面构建起完整的能力体系。 "京东在AI ...

"AI医院1.0"汇聚了包括AI医生大为、小微，以及AI营养师小晶、AI药师小方等在内的十多类专业服务 智能体，以及超过1000名来自全国400余家三甲医院的专家医生智能体。这些智能体依托"京医千询"大 模型，能为用户提供7*24小时不间断的在线导诊、问诊、报告解读、健康管理等服务，以其温暖共情的 交互体验赢得了用户的广泛好评，2025年至今已累计服务用户超1.5亿，服务满意度高达97%。 依托"京医千询2.0"技术底座，京东健康全新推出面向真实医疗场景的医生Agent。它实现了行业首创的 完全自由问诊对话，支持自由多问多答，能基于专业性和用户诉求灵活回应。同时，基于临床真实世 界，借助海量真实问诊数据的高仿真迭代系统，以及医学专家隐性经验的持续输入与对齐，精准模拟真 实医疗场景。此外，还有密切的人机协同，搭配百余种专业和业务工具，比如医学计算、多模态辅助， 以及处方、检验单、随访等业务相关工具，为医疗服务提供有力支持，助力更高效、精准的诊疗。 会上，京东健康发布全新的"AI医院1.0"，在行业内首次实现AI驱动的"医检诊药"服务闭环，即从导诊 到问诊、从检查到买药，真正让AI实现对用户实际需求的准确理解和主 ...

西风 一水 发自 凹非寺 量子位 | 公众号 QbitAI 超越DeepMind的Langfun、抱抱脸的Smolagent等，国产智能体位列全球智能体第一梯队！ GitHub明星开源项目、行业首个 100%开源的企业级智能体JoyAgent ，迎来重要更新—— 进一步开源DataA gent和DCP数据治 理模块，加速企业级落地。 升级后的JoyAgent，在全球权威智能体评测榜单 GAIA， Validation集准确率77%，Test 集准确率超67 % ， 相较于此前成绩实现稳步提 升。 该项目在GitH ub的热度也顺势走高， Star数迅速过万 ，截至发稿已达10.1k。 | | ( jd-opensource / joyagent-jdgenie Public template | | | | | | | A Notifications | y Fork 1k | Star 10.1k | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | <> Code | · Issues 107 I Pull requ ...

9月25日，主题为"Enjoy AI"的JDDiscovery-2025京东全球科技探索者大会在北京举行。京东在会上发布 了AI 全景图，系统展示了AI整体战略布局，未来三年将持续投入，带动各个产业形成万亿规模的人工 智能生态。 △京东集团 SEC副主席、CEO许冉演讲 △京东AI 全景图 物流全新升级的物流超脑大模型2.0全面走向多模态，标志着物流的"超脑"正在从辅助决策到具身执 行。通过多个智能设备协作，大幅提升物流供应链作业效率，标准化操作水平提升15%，人机协同作业 效率提升20%；首次发布的异狼具身智能机械臂，是自研"狼族系列"的又一项突破性创新，可解决海量 非标包裹的自动化码垛难题，极大提高了物流仓储效率和准确性。目前， "狼族" 智能设备全球部署仓 库超 500 个，覆盖存储、搬运、分拣全链路，标志着新一代物流智能形态的代际跨越。 零售首次对外公布电商创新AI架构体系Oxygen。该体系依托Joy AI大模型打造丰富的系统能力和多元化 智能体，通过AI技术构建电商购物、供应链管理等智能应用场景，系统展现了AI技术在京东电商业务 中的应用，描摹出京东零售面向未来的AI电商蓝图。同时将于10月在京东 ...

近期高度活跃的京东，又对AI"下手了"。 在今天举行的JDD大会上，京东一口气发布了三大AI产品、三个升级的深度应用平台，以及京东自研AI在四大场景的应用成果。同时京东集团SEC副主席、 京东集团CEO许冉宣布，未来三年将持续投入人工智能发展，带动万亿规模的人工智能生态，并向具身智能等人工智能企业开放场景与训练数据。 京东这次对自研AI能力的集中展示，事实上为观察当前互联网行业的AI应用落地，提供了一个差异化样本。京东采用的"场景驱动"策略已见成效：先在自身 供应链体系中验证AI应用的实际价值，再将经过实战检验的能力做了系统化输出。 这一路径选择背后，既代表着京东对AI战略意义的本质理解，也呼应了当前AI从强调技术研发转向注重实际应用价值的行业大趋势。京东的实践，或许能 为AI深度应用提供可参照的关键要素和可能路径。 京东的AI战略， 从一开始就脱离了炫参数和炫技能的层面。 自用打磨，AI重塑供应链 从JDD大会的发布来看，京东在"关起门来"自研自用AI的这段时间，已经实现了自研AI能力跑通核心业务场景的成果。 以供应链的枢纽环节物流场景为例，京东在自建物流体系中已经完成了大模型的深度嵌入。经过升级后的物流超 ...

今天，在京东全球科技探索者大会（JDD）上，京东宣布，未来三年，将持续投入，带动形成万亿规模的人工智能生态。 这次大会，京东首次系统展示了AI全景布局，不仅升级"JoyAI"大模型品牌，还一口气发布三款C端产品、四大场景应用和三大深度平台，意图把供应链 优势延伸至用户端。 受此消息带动，京东集团港股股价盘中涨幅近7%。 9月25日，京东宣布：未来三年，将持续投入，带动形成万亿规模的人工智能生态 三款产品全面押注AI新交互 同时，京东正式推出行业首个"附身智能平台"JoyInside 2.0，强调全龄陪伴、拟人交互和言行一致三大能力，为机器人、AI玩具、智能眼镜等终端"装上大 脑"。目前已有超过30家硬件品牌接入，相关产品已上线京东。 JoyInside2.0已接入超30家品牌 即将上线的京犀App，被定位为"下一代购物与生活服务超级入口"，不只"听懂人话"，更理解商品，意图实现真正意义上的超个性化推荐。据透露，AI购 物、AI试衣等体验后续将从主站迁入这款独立App。 "他她它"App则搭载新版JoyAI大模型，内置一位什么都能聊的"万能博士"智能体，既能查天气、点外卖、聊财经，还可根据用户性格切换对话风格。 ...

今天，在京东全球科技探索者大会（JDD）上，京东宣布，未来三年，将持续投入，带动形成万亿规模 的人工智能生态。 这次大会，京东首次系统展示了AI全景布局，不仅升级"JoyAI"大模型品牌，还一口气发布三款C端产 品、四大场景应用和三大深度平台，意图把供应链优势延伸至用户端。 受此消息带动，京东集团港股股价盘中涨幅近7%。 即将上线的京犀App，被定位为"下一代购物与生活服务超级入口"，不只"听懂人话"，更理 解商品，意图实现真正意义上的超个性化推荐。据透露，AI购物、AI试衣等体验后续将从主站 迁入这款独立App。 "他她它"App则搭载新版JoyAI大模型，内置一位什么都能聊的"万能博士"智能体，既能查天 气、点外卖、聊财经，还可根据用户性格切换对话风格。同时，推出"AI圈子"，借智能体互 连帮用户寻找同频伙伴，甚至可嵌入多类硬件，实现IP联名和"记忆一致"的跨设备体验。 同时，京东正式推出行业首个"附身智能平台"JoyInside 2.0，强调全龄陪伴、拟人交互和言 行一致三大能力，为机器人、AI玩具、智能眼镜等终端"装上大脑"。目前已有超过30家硬件 品牌接入，相关产品已上线京东。 从模型升级到生态扩容 ...

Core Insights - The article highlights that JD Health has been recognized for its leading AI capabilities and applications in the healthcare sector, being included in the "2025 Sustainable Innovation Case Recommendation List" at the World Artificial Intelligence Conference [1][3]. Group 1: Recognition and Achievements - JD Health was selected for its significant contributions to sustainable innovation in the healthcare industry, alongside over 30 other renowned companies [3]. - The company has developed a series of AI medical products, including the "JD Medical Inquiry" model, which is the first fully open-source vertical model in China's medical industry [3][4]. Group 2: AI Model Development - The "JD Medical Inquiry" model is continuously upgraded, focusing on a "three engines + four models" architecture to enhance its capabilities in complex disease areas [4][10]. - The upgraded version "JD Medical Inquiry 2.0" integrates multi-source clinical information to provide reliable decision support for doctors, particularly in specialized disease treatment [8][9]. Group 3: Application and Impact - JD Health's AI applications have rapidly expanded across various healthcare scenarios, serving over 30 million users within six months of launching the "Kangkang" health assistant [11]. - The AI-driven services have lowered the barriers for users to access online medical consultations, allowing them to receive scientific advice without needing to visit a doctor first [12][13]. Group 4: Future Directions - JD Health aims to further enhance its AI capabilities by developing deep reasoning and multi-modal models to provide personalized and precise medical consultations [10]. - The company plans to expand its AI doctor program, aiming to cover more specialized areas and enhance the integration of online and offline healthcare services [11][13].

Core Insights - The article highlights the resurgence of medical AI at the WAIC, with major tech companies and startups re-engaging in the sector, indicating a shift in focus towards healthcare applications [1][2] Group 1: Medical AI Landscape - Major tech companies like Tencent, Alibaba, and ByteDance are refocusing on medical AI as a core business area, showcasing advanced AI solutions at WAIC [1] - Key topics discussed include AI drug development, clinical applications, and the democratization of healthcare, with participation from pharmaceutical and medical device companies [1] Group 2: Pathways of Medical AI Development - Pathway One: Fragmented AI is beginning to systematically empower healthcare, moving from isolated applications to integrated solutions that address broader healthcare needs [2][3] - The emergence of intelligent agents allows for proactive, goal-driven interactions in healthcare, enhancing the efficiency of medical processes [3][4] Group 3: Intelligent Agents and User Engagement - Intelligent agents like Tencent's "Health Management Assistant" integrate various health tools, enabling continuous user engagement and proactive health management [3][4] - Startups are also leveraging AI to enhance low-digitalization scenarios, such as simulating patient interactions for training purposes [5] Group 4: Clinical Applications and Specialized Models - Pathway Two: The focus is shifting from general applications to specialized clinical models, with companies like JD Health leading the way in developing tailored AI solutions for specific medical fields [6][7] - JD Health's "京医千询2.0" aims to enhance AI capabilities in clinical settings, focusing on evidence-based data and interactive simulations [6][7] Group 5: AI in Medical Imaging - Union Medical is noted for its advancements in medical imaging AI, showcasing a unique intelligent agent capable of detecting multiple chest abnormalities with high accuracy [8][9] - The intelligent agent demonstrates significant efficiency improvements in diagnostic processes, highlighting the potential of AI in enhancing clinical workflows [9] Group 6: Standardization and Regulation - The need for a comprehensive regulatory framework for medical AI is emphasized, with initiatives like the "AI Doctor" standard being introduced to ensure quality and safety in AI applications [11][12] - The establishment of standards is seen as a crucial step towards systematic and regulated development in the medical AI sector [12] Group 7: Future Outlook for Medical AI - Despite challenges in monetization and ethical considerations, there is optimism about the potential for medical AI to become an integral part of the healthcare ecosystem [13][14] - Collaborative efforts among government, healthcare institutions, and tech companies are paving the way for a more robust medical AI landscape, with the possibility of achieving viable business models in the near future [14]

Core Insights - JD Health has launched "Jingyi Qianxun 2.0" focusing on advancing medical AI from general services to specialized fields through technological innovation and ecosystem openness [1][2] - The architecture consists of "Three Engines" and "Four Models" aimed at enhancing capabilities and ensuring reliable medical reasoning [1][2] - Significant breakthroughs have been achieved in human-like dialogue, credible reasoning, and multi-modal medical data analysis [1][2] Group 1 - The "Three Engines" include evidence-based data engine, clinical case engine, and patient-doctor interaction simulation engine [1] - The "Four Models" consist of general practitioner model, specialist model, health agent, and imaging model [1] - "Jingyi Qianxun 2.0" emphasizes human-like dialogue that simulates real doctor-patient interactions and provides evidence-based recommendations [1] Group 2 - The "medical multi-modal" approach integrates text, images, and test data for comprehensive analysis to support precise diagnosis [2] - JD Health's first large model product for hospitals, "JD Zhuoyi," has been implemented in several medical institutions [2] - The focus on "Three Engines and Four Models" aims to accelerate the transformation of medical intelligence rather than just developing larger models [2][3]