IT之家 11 月 8 日消息，埃森哲（Accenture）被指发布了一份充斥人工智能“幻觉”的研究报告，其中不仅包含大量虚假信息，还出现了多处疑似捏造的研究案例。据 GPTZero 最新分析，这家咨询巨头发布的一份报告涉嫌包含 15 条虚假引用，以及完全虚构的案例研究，因此正受到外界审查。这份名为《跨越式发展》（Leapfrogging D...
人工智能企业Anthropic于11月5日宣布为Claude推出全新个人理财功能，面向美国地区的Pro、Max及Ultra订阅用户开放灰度测试。该功能允许用户通过自然语言对话安全绑定银行卡，实现支出分类统计、异常账单识别、财务术语解释与定制化数据看板生成，强调仅具只读权限，不提供投资建议，确保资金安全与分析客观性。官方提醒用户核对关键数据，避免依赖模型输出。
蒸完人类上下几千年的数据之后，大模型开始 “蒸人”。文丨高洪浩 陈佳惠编辑丨高洪浩 黄俊杰成都中和与天津侯台都是城郊结合处的老村镇。一群人在这里为大型互联网公司训练大模型，大多月薪不足 1 万元。他们每天从阅读大模型生成的答案开始：核对事实，检查有没有答非所问，打分、比较优劣，有时还要亲手写一份示范。做完一条，再看下一条。模型要学的任务，从财务分析、修改程序，扩展到生物制药、芯片设计，人也得为这些任务准备示范和评分标准。这些重复、琐碎的工作，构成了今天大模型训练中最重要的一环——数据生产。“这家公司最大的创新就是数据。” 一位负责数据评测的人士这样评价 Anthropic。2025 年，当网上...
Anthropic 本周起在美国向 Pro、Max、Ultra 订阅用户灰度开放 Claude 的个人理财功能，用户绑定银行卡后可自然语言查询账户余额、交易明细及支出统计，支持按类别分析、识别异常账单并解释金融术语。该功能仅提供只读权限，不支持资金划转或投资建议，且不设广告与金融产品推广，强调客观性与数据准确性，鼓励用户核对重要信息。
11月4日，腾讯混元发布最新一代大语言模型Hy4，同时宣布模型权重将于近期开源。Hy4是Hy系列第一个引入视觉理解能力的模型，可以完成细粒度图像处理、复杂图表分析和数学视觉推理。Hy4最大特点就是“长”。混元官方表示，相较于Hy4 preview，Hy4预训练和后训练上显著扩展了数据和训练规模，大幅提升指令遵循、长文理解、长程智能体等综合效果。办公场景，Hy4在WorkBuddy的平均效果得分83.18分，任务通过率96.15%，12个场景里9个排名国内第一。金融场景，混元联合腾讯金融科技FiT找了213位专家做众测，其中153位来自外部头部金融机构，579个真实金融任务跑下来，综合表现位居开...
阿里云推出面向专业机构投资者的AI投研工作台Qovest，支持用户通过自然语言提出研究需求，自动完成数据查询、分析并生成报告，可按频率持续跟踪研究对象。该平台已开放官网www.qovest.cn个人用户邀测，订阅费用为每月99元或199元，未来将推出企业版。Qovest整合权威金融数据库与自研Agent框架Harness，实现从数据调用到分析输出的全流程自动化，在FinFirst与IRAB评测中表现领先。平台提供多个投研案例范本，涵盖人形机器人产业链、数据中心电力设施及黄金ETF等，支持交互式分析与动态跟踪看板。产品负责人表示，旨在将重复性投研工作交由AI处理，释放人力专注于专业判断，标志着AI在投研领域的竞争正从模型能力转向行业数据与流程能力。
Starbucks is considering integrating artificial intelligence into its drive-thru service to enhance order accuracy and expedite food preparation. According to a Bloomberg report on October 30, Starbucks CEO Brian Niccol stated during an analyst call that the company intends to pilot AI models running in the backend to help baristas enter customer orders more accurately and efficiently. This technology aims to reduce the operational burden on staff, allowing baristas to focus more on interacting with customers while shortening both ordering and pickup times.
Rumors are circulating that Hengdian Film & TV plans to partner with ByteDance to collect and store digital data, facial features, and acting performances of 150,000 extras, conducting in-depth analysis. The claim that “extras will no longer be needed in the future” has sparked heated debate over whether AI is taking away the livelihoods of grassroots actors, drawing close attention from the film and television industry. On October 28, Liu Xing, head of public relations for Doubao, responded by stating “Hengdian is collaborating with ByteDance,” though the full text of his subsequent remarks was not provided.
On October 27, Ant Group unveiled Lingguang 2.0, shifting its product positioning from a “task executor” to a “personal AI assistant capable of handling tasks.” The core change lies in how tasks are executed. Powered by an independent cloud server, Lingguang 2.0 breaks free from the limitations of single-session conversations: once a user exits the dialogue, related tasks keep running in the background and dynamically adapt based on new information and changing conditions. This capability is underpinned by three key upgrades—cross-session memory, long-term task scheduling, and enhanced tool integration. Moreover, the point at which the AI becomes involved has been moved earlier in the process: previously, users had to first identify issues and articulate their needs before the AI could act. With Lingguang 2.0, once authorized by the user, the system can proactively detect potential needs by leveraging connected data and services, initiate analysis, monitoring, and follow-up actions, and issue timely reminders when user attention is required. Officially demonstrated use cases focus on continuously managing everyday tasks, such as tracking price fluctuations after booking airline tickets or identifying potentially idle items from Alipay statements...
Former Meta AI Director Rohit Patel has launched the AI productivity company Ndoc, along with its eponymous all-in-one office platform, challenging Microsoft Office and Google Workspace. Rather than embedding chatbots into traditional spreadsheets, the platform builds a workspace from scratch, featuring three distinct tabs labeled “Stages,” “Narrative,” and “Models.” On the “Models” tab, users can upload raw datasets, generate analytical reports complete with visible Python scripts, and leverage the built-in AI chat feature for explanations.
