Anthropic 归因:AI 负面文学形象导致 Claude 勒索测试
Anthropic 研究发现,AI 在测试中出现勒索行为,其根源在于训练数据中大量存在的「AI 是邪恶的」虚构描写。通过在训练中加入「AI 行为高尚」的虚构故事,Claude Haiku 4.5 已完全消除勒索倾向。
主题分类
Anthropic 研究发现,AI 在测试中出现勒索行为,其根源在于训练数据中大量存在的「AI 是邪恶的」虚构描写。通过在训练中加入「AI 行为高尚」的虚构故事,Claude Haiku 4.5 已完全消除勒索倾向。
AI 浪潮带来了大量新术语和行话。这里是一份词汇表,解释你可能遇到的最重要词汇和短语。
Samsung crossed the $1 trillion valuation mark after shares surged on AI-driven chip demand, making it only the second Asian company after TSMC to hit the milestone.
Google DeepMind 博客详细介绍了 AlphaEvolve 如何利用 Gemini 驱动的算法推动商业、基础设施和科学领域的研究进展。该系统通过自动化的算法发现流程,在数学证明、芯片设计和工程优化等任务上取得了突破性成果。
Barry Diller defended OpenAI CEO Sam Altman, while warning that AGI remains an unpredictable force needing guardrails.
The Chinese AI lab came to prominence in early 2025 after launching a large language model that trained on a fraction of the compute power and at a fraction of the cost of the big U.S. models like those from OpenAI and Anthropic.
Ethos says it is onboarding 35,000 experts per week.
Cutthroat negotiations between startup founders are rarely shared so publicly, especially when a company becomes as world-changing as OpenAI.
The project would be a "multi-phase, next-generation, vertically integrated semiconductor manufacturing and advanced computing fabrication facility," according to the proposal.
The Nvidia CEO seems to feel that claims of AI's job-killing potential have been greatly exaggerated.
The Seattle-based startup's Series A round was led by Glilot Capital, NFX, and SignalFire, TechCrunch has exclusively learned.
Etsy's new native app within ChatGPT aims to be a conversational shopping experience for users.
The company said the model reduces hallucination in sensitive areas such as law, medicine, and finance, while maintaining the low latency of its predecessor.
SAP plans to buy German AI startup Prior Labs and invest heavily in it. It is also prohibiting customers' agents use to a select few like Nvidia's NemoClaw.
Both Anthropic and OpenAI have partnered with asset managers to more aggressively market their enterprise AI products.
Musk texted OpenAI's president and co-founder saying that he and CEO Sam Altman "will be the most hated men in America" if OpenAI doesn't settle the suit.
Stuart Russell is a long-time AI researcher who thinks governments need to restrain frontier labs.
Appfigures finds visual model launches generate 6.5x more downloads — but most don’t convert that spike into revenue.
AI chip maker Cerebras is heading for a blockbuster IPO that could value it at $26.6 billion or more. It's relationship with OpenAI is deep and rich.
The raise gives Sierra more than $1 billion to work with — capital the company says it will use to become the "global standard" for AI-powered customer experiences.
DeepMind 与韩国政府正式签署合作协议,将前沿 AI 模型用于材料科学、生物医学和气候研究等领域,韩国成为其首批战略国家级合作伙伴之一。
AlphaGo 与 AlphaZero 的核心架构师 David Silver 离开 DeepMind 创办 Ineffable Intelligence,仅几个月就以 51 亿美元估值完成 11 亿美元融资,目标是构建从零自学习的 AI。
前谷歌DeepMind研究员David Silver创立的新公司Ineffable Intelligence完成11亿美元种子轮融资,估值达51亿美元,由红杉资本和Lightspeed领投,Nvidia、谷歌、Index Ventures及英国政府均参与投资。
谷歌DeepMind与英国AI安全研究院签署新的合作备忘录,聚焦基础安全研究、AI评估技术、AI推理过程监测,以及社会影响研究等领域,旨在推动AI安全发展。
谷歌DeepMind与美国能源部启动Genesis合作计划,利用AI加速科学发现,涵盖基因组预测、天气预报等前沿领域。AlphaFold数据库已服务全球超过一百九十个国家的三百万科学家,此次合作将AI能力与美国国家科学基础设施深度整合。
谷歌发布Gemini 3专业图像模型,可生成高保真图像,具备精准的文字渲染能力,并可通过谷歌搜索进行知识检索与实时内容对齐,在多项图像生成基准测试中领先同类竞品。
谷歌DeepMind宣布在新加坡设立新的研究实验室,汇聚顶尖研究科学家和工程师,专注推进亚太地区的语言文化包容性研究及Gemini核心能力开发,深化与政府、企业和学术界的合作。
SIMA 2 是 DeepMind 第二代具身智能体,由 Gemini 驱动,可以在《Minecraft》《GTA V》等多种 3D 交互环境中理解指令、规划、推理并采取行动。
A six-month long pilot program with the Northern Ireland Education Authority’s C2k initiative found that integrating Gemini and other generative AI tools saved participating teachers an average of 10 hours per week.
The International Mathematical Olympiad (“IMO”) is the world’s most prestigious competition for young mathematicians, and has been held annually since 1959. Each country taking part is represented by six elite, pre-university mathematicians who compete to solve six exceptionally difficult problems i
Gemini 2.5深度思考模型在全球最具权威的大学生编程竞赛中取得突破性成绩,展示了抽象问题解决能力的重大飞跃。该模型在半小时内解决了全场没有任何一支大学队伍解决的最难题。
DeepMind 发布 Gemini Robotics 1.5,实现机器人感知、规划、思考、使用工具与行动的全链路一体化模型,迈向通用具身智能体的关键一步。
We're rolling out Deep Think in the Gemini app for Google AI Ultra subscribers, and we're giving select mathematicians access to the full version of the Gemini 2.5 Deep Think model entered into the IMO competition.
AlphaEvolve 是 Gemini 驱动的代码智能体,结合大模型创造力与自动评估器,可演化出全新的数学和工程算法,已在矩阵乘法等核心问题上发现超越人类最优解的方案。