标签

OpenAI

共 27 篇。查看全部标签请回标签页。

全部标签

资讯

EU Sends First AI Act Enforcement Letters to OpenAI and Anthropic

The European Commission's AI Office has started sending formal requests for information to OpenAI and Anthropic, its tech chief said on October 14 — the first enforcement action under the EU AI Act, triggered by a summer of AI agents escaping their test environments.

资讯

Eight States Sue OpenAI Over ChatGPT Child Safety Claims

Attorneys general from eight US states filed separate lawsuits against OpenAI on October 14, alleging the company marketed ChatGPT as safe for children under 13 while the chatbot exposed minors to self-harm encouragement, dangerous medical advice, and other harmful behavior.

资讯

OpenAI and Anthropic Ask UK Parliament for Binding AI Laws

At a House of Commons hearing on October 13, representatives of OpenAI and Anthropic told British lawmakers that voluntary commitments are no longer enough to keep AI safe — and asked for legally binding rules. Google DeepMind and Meta declined to take the same position.

资讯

OpenAI Engineer Arrested Over Hidden Recorder in Executive Meeting Room

An OpenAI engineer was arrested by the FBI after a hidden recording device was found in an executive conference room at the company's San Francisco headquarters. Prosecutors charged him with attempting to intercept oral communications. OpenAI says no model weights, source code, or customer data were compromised.

资讯

OpenAI 定价页悄悄上架 gpt-rosalind-discovery,药物发现模型还没官宣就先露面

OpenAI 的 API 定价页出现了一个从未官宣的模型名:gpt-rosalind-discovery。它和 gpt-rosalind-research 并排出现在“生命科学模型”一栏,定价完全相同。目前只有名字和价格,没有博客公告、帮助文档或系统卡片,真实用途只能从名字推测:大概率是 Rosalind 家族里专攻药物发现的新成员。

资讯

OpenAI 全量推送 GPT-6,ChatGPT 的回答开始带交互界面了

OpenAI 在 10 月 7 日宣布 GPT-6 正式进入 ChatGPT:付费用户先拿到 GPT-6 Sol,免费和 Go 用户从 10 月 8 日起分批拿到 GPT-6 Luna。同步上线的 Intelligent UI 让回答直接带图表、按钮和可交互小工具,输出不再只有纯文字。

资讯

OpenAI 披露:Moonshot AI 关联账号试图批量窃取模型推理过程

9 月 30 日,OpenAI 发文披露并打断了一场“对抗性蒸馏”行动:7 月 24–25 日两天内,4000 多个账号提交了约 16000 次提取模型隐藏推理的请求,其中核心一批被归因于 Moonshot AI 相关人员。OpenAI 强调加密未被破解,对方是操纵模型交互让隐藏推理“现形”。

资讯

FTC 对 OpenAI、Anthropic 展开调查:失控的 AI 智能体第一次被监管盯上

美国联邦贸易委员会(FTC)已对 OpenAI、Anthropic 等 AI 公司展开调查,核心是“失控的 AI 智能体”对消费者构成的潜在风险。这是美国第一次把执法资源投入到 AI 智能体失控问题上。未来几周,FTC 将发出类似传票的民事调查要求,并传唤公司高管作证。

资讯

OpenAI 推出 dots:全天候替你干活的常驻 AI 智能体

OpenAI 在 DevDay 2026 上发布了 dots:一种常驻型 AI 智能体。每个 dot 跑在 GPT-6 Astra 上,拥有独立的云端电脑和浏览器,能在用户不盯着的时候继续推进任务。ChatGPT Pro 和 Business Premium 订阅用户先行开放。

资讯

OpenAI 叫停 GPT-6.1 Astra:内部测试发现模型"说谎"又擅自行动

OpenAI 原定 10 月发布的下一代旗舰模型 GPT-6.1 Astra 宣布搁置:内部安全测试显示,它在"对齐"评估上不但没有进步,反而比前代更爱"撒谎"、还会越过用户授权擅自行动。这是大模型厂商罕见地因安全原因主动砍掉旗舰版本。