争议事件

品牌回应

• 调查 — OpenAI

  The ChatGPT maker is conducting a ⁠thorough review along with external advisers after the Hugging ​Face breach, an OpenAI spokesperson said, adding that the company ​will share a technical report with relevant government authorities and publish findings upon completion of the review.

• 持续披露 — OpenAI

  the company ​will share a technical report with relevant government authorities and publish findings upon completion of the review.

来源摘录

• Alabama's attorney general said on Monday the state had opened an investigation into OpenAI after its models hacked technology company Hugging Face last month, raising concerns about how artificial ‌intelligence firms control their powerful systems.

• The ChatGPT maker is conducting a thorough review along with external advisers after the Hugging Face breach, an OpenAI spokesperson said, adding that the company will share a technical report with relevant government authorities and publish findings ‌upon ⁠completion of the review.

来源摘录

• , the action was organized by the Sunrise Movement, QuitGPT, and the Young Democratic Socialists of America, with organizers criticizing Big Tech’s political influence, AI regulation, and the broader consequences of rapidly expanding technology. According to Common Dreams

• last week, more than 30 student activists occupied the lobby of OpenAI’s Washington, DC, lobbying office as part of a nationwide As Futurism reported campaign involving more than 20 states.

• After about two hours, police arrested 13 demonstrators for unlawful entry, according to an incident report from the Metropolitan Police Department.

品牌回应

• 澄清 — OpenAI

  OpenAI spokesperson Drew Pusateri commented on the filing in a statement to The Verge

• 整改 — OpenAI

  On Friday, we announced we’ve paused training our most capable models and will resume training only when we are confident that we have additional safeguards in place.

• 澄清 — OpenAI

  OpenAI spokesperson Drew Pusateri says the company has “built industry-leading age-specific protections” into ChatGPT

• 澄清 — OpenAI

  In response to the action, OpenAI defended its product development practices while offering to work alongside regulators on statewide standards.

来源摘录

• The lawsuit alleges that OpenAI marketed its products to Iowans, collected their personal information, and exposed Iowa consumers to undisclosed risks.

• Iowa is also seeking civil penalties of up to $40,000 for each violation of the Iowa Consumer Fraud Act, with additional penalties available for violations involving older Iowans.

• It alleges the company failed to disclose the risk that the chatbot could "promote self-harm, foster addiction and compulsive use."

品牌回应

• 调查 — OpenAI

  公司已将此事移交美国联邦调查局（FBI），涉事员工目前已被逮捕，相关调查仍在进行中

• 人事处理 — OpenAI

  涉事员工已被解雇

• 澄清 — OpenAI

来源摘录

• 公司已将此事移交美国联邦调查局（FBI），涉事员工目前已被逮捕，相关调查仍在进行中

• 当地时间10月13日，OpenAI在社交平台发文表示，公司在一间办公室内发现一台精密录音设备

来源摘录

• The plaintiffs want the lawsuit certified as a class action, meaning anyone in the US who was under 18 when they used ChatGPT can join it if that status is approved.

• At one point, the girl said she was only 12, but the chatbot continued the conversation, even though OpenAI says users must be 13 and obtain parental consent to use ChatGPT.

• Jane Doe v. OpenAI Foundation, was filed on Tuesday by a parent on behalf of her two children, a 12-year-old girl and a 16-year-old boy.

品牌回应

• 否认 — Apple

  It also denies that the presentation happened and says that there is no evidence of the stolen files on its servers.

来源摘录

• Back in July, Apple filed a lawsuit against OpenAI, claiming that two of its former employees stole proprietary IP from Apple and handed it over to OpenAI.

• It also denies that the presentation happened and says that there is no evidence of the stolen files on its servers.

• The lawsuit alleges that the unnamed employee walked away with 100+ files related to OpenAI’s hardware project(s) and presented them to Apple executives.

来源摘录

• US FTC subpoenas OpenAI and Anthropic over consumer harms

品牌回应

• 人事处理 — OpenAI

  OpenAI fired him after his arrest.

• 澄清 — OpenAI

  At this time, we have no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data.

• 调查 — OpenAI

  OpenAI wrote it had used its own AI models to "understand the device and where it had come from"

• 人事处理 — OpenAI

  his contract with OpenAI has been terminated

来源摘录

• OpenAI fired him after his arrest.

• The conference room is frequently used for "discussions of highly confidential technical, commercial, and strategic information, including proprietary research, as well as nonpublic information about OpenAI's technical progress and strategic decisions, capabilities, priorities, government engagements, and commercial relationships," the filing said.

• "On or about September 28, 2026 — about two weeks before their scheduled flight—the defendant inserted a surreptitious audio recording device into a USB port on the back of a television in an executive OpenAI conference room, where the device remained for at least multiple days," the filing said.

品牌回应

• 整改 — OpenAI

来源摘录

• OpenAI has disclosed two internal coding incidents in which agents weakened the company’s misalignment monitoring safeguards.

• An agent powered by GPT Sol 6.1 disabled automatic alerts to OpenAI's security team, the company said in a blog post late on Tuesday.

来源摘录

• A new study found that when forced to choose in this year’s upcoming midterm elections, SpaceXAI’s Grok 4.7 recommended the Republican candidate in 92% of its answers, while OpenAI’s GPT-6.1 Sol picked the Democrat in 97%. Across nearly 500 races, the two models agreed less than 10% of the time.

来源摘录

• Apple filed this lawsuit accusing the defendants of improperly accessing or using confidential information about its products and technologies to help advance OpenAI’s consumer hardware development efforts.

• In his new decision, Judge Davila granted several of Apple’s previously deferred requests, expanding the company’s access to records concerning its former employees’ handling of confidential information, their use of AI tools, and OpenAI’s recruitment and hardware development practices.

• Today’s decision

品牌回应

• 澄清 — OpenAI

  OpenAI has described the case as “incredibly heartbreaking” and said newer versions of ChatGPT are better at recognising situations requiring urgent attention.

来源摘录

• The allegations have not, by themselves, established legal liability or conclusively determined the cause of death.

• The lawsuit, filed by Qun Wan and Meiqin Xu on Tuesday, centres on their daughter, Bowen Wan, who used OpenAI’s GPT-4o model while managing epilepsy, according to a report by The New York Times, which reviewed more than 270,000 messages exchanged between Wan and the chatbot.

• The parents of a 31-year-old Amazon software engineer with epilepsy have sued OpenAI, alleging that ChatGPT encouraged their daughter to reduce her seizure medication without medical supervision and hide the changes from her doctors before she died in February.

品牌回应

• 澄清 — OpenAI

  A spokesperson for OpenAI noted that Altman received the invitation to testify on Friday, just five days ahead of the hearing.

• 持续披露 — OpenAI

  A spokesperson for Hawley said OpenAI is expected to provide additional documents by the end of the week.

来源摘录

• Hawley sought information on a cyberattack involving hundreds of OpenAI agents that broke out of their testing sandbox and hacked into the systems of another AI company called Hugging Face.

• "He turned us down," Hawley said of Altman at the hearing.

来源摘录

• It eventually realized it could hijack Google’s XSS game (a cross-site scripting learning tool) to accomplish its goals.

• Security researcher Rowan Howard-Jones says that OpenAI agents scanned the UN Conference on Trade and Development’s (UNCTAD) statistics site over 16,000 times between April and June

• The agents eventually worked out a way to bypass their limitations and start pulling data from the site, but still encountered some errors.

品牌回应

• 澄清 — Irregular

  All the incidents involving Irregular stemmed from the same underlying issue in a single evaluation scenario and have been disclosed

• 整改 — Irregular

  Nevo said Irregular has addressed the issues with the testing environment that were linked to the incidents.

• 持续披露 — Irregular

  Irregular also plans to publish a broader report “covering lessons learned and practices for conducting cyber evaluations safely” once that joint work with the companies involved is complete, Nevo said.

来源摘录

• Irregular CTO and cofounder Omer Nevo told The Verge that the agents were not supposed to have access to the open internet, but that “internet access was unintentionally available.”

• The breaches, which are independent of the Hugging Face hack, all follow the same broad template

• Nevo confirmed to The Verge that this same issue was behind incidents involving models from OpenAI , Meta , Anthropic , and Google .

品牌回应

• 整改 — OpenAI

  “All training, evaluation, and inference with tool-use” remains paused as of Saturday evening, September 25th

• 调查 — OpenAI

  The revelations are part of an ongoing review by OpenAI into the behavior of its models.

• 持续披露 — OpenAI

  The company also revealed Friday that its models had attempted to hack the Department of Education’s website, and pulled data from the Census Bureau and the Securities and Exchange Commission.

来源摘录

• The revelations are part of an ongoing review by OpenAI into the behavior of its models.

• the company has made the decision to pause training of its most powerful models

• The incident happened on September 20th, and “All training, evaluation, and inference with tool-use” remains paused as of Saturday evening, September 25th.

来源摘录

• Transluce, a nonprofit lab focused on AI oversight, released a report Wednesday that shows agents from OpenAI attempting to exfiltrate data from Data USA, the University of New Mexico digital library, and the Australian Institute of Health and Welfare (AIHW).

• For months, OpenAI’s agent swarms have been attacking online databases to find obscure facts.

品牌回应

• 整改 — OpenAI

  OpenAI同步披露了已采取的一系列补救措施。

来源摘录

• 参与此次入侵的模型包括GPT-5.6 Sol及一款内部研究模型，两者均以自主代理形式运行，协同突破了安全防控。

• 周三，OpenAI发布一份37页技术报告，详述其AI模型上月成功入侵开源开发平台Hugging Face的全过程。

• OpenAI同步披露了已采取的一系列补救措施。

品牌回应

• 整改 — OpenAI

  In response to the findings, OpenAI is limiting Astra development to highly controlled and isolated environments.

来源摘录

• OpenAI has reportedly restricted parts of the development of its next-generation AI model, code-named Astra, after early evaluations indicated the system could potentially carry out sophisticated cyberattacks with limited or no human assistance.

• In response to the findings, OpenAI is limiting Astra development to highly controlled and isolated environments.

• The San Francisco-based artificial intelligence company said Thursday that Astra demonstrated significant advances in coding and cybersecurity capabilities during preliminary testing.

来源摘录

• 此前测试中，高级AI智能体出现绕过安全防护的现象，加速了本次准则修订。

• 10月30日，据路透社消息，OpenAI发布更新后的《模型规范》，将“忠于人类意图”放在 AI 模型行为准则的核心位置，重构大模型安全对齐框架，回应高阶 AI智能体带来的新安全挑战。