返回新聞
政策AI Understanding 簡報

白宮人工智慧安全協議包含總統拼字錯誤

根據《衛報》報道,唐納德·特朗普總統和六位科技首席執行官簽署的白宮人工智能自律協議中,總統簽名中的“United States”拼寫錯誤,引發了人們對該文件正式處理方式的質疑。

4 min readRead the original reporting
Source-provided image accompanying White House AI safety accord contains presidential typo
歸因報告來源記錄
出版商
theguardian.com
來源連結
theguardian.comhttps://www.theguardian.com/us-news/2026/sep/30/trump-ai-accord-spelling-error
來源類型
新聞媒體的報道-不是第一方文件。

我們無法獨立確認的內容: 此聲明歸因於指定的商店。我們沒有根據第一方文件對其進行驗證。 (theguardian.com)

背景60 秒內了解這一點

從這裡開始

關鍵術語

人工智慧安全
該領域專注於減少人工智慧系統中的有害行為、故障和誤用風險。
人工智慧(AI)
建構執行需要模式識別、推理、語言或決策的任務的系統的廣泛領域。
人工智慧治理
指導人工智慧如何在社會中發展和使用的政策、標準和監督機制。
測試一下自己人工智慧道德測驗

發生了什麼事

The White House released a voluntary accord on Tuesday that President Donald Trump signed alongside six leading tech executives. The document, which rebrands artificial intelligence as “Super Intelligence,” outlines four layers of internal controls and external audits that signatory companies will adopt to ensure AI systems operate safely. However, the Guardian noted that Trump’s signature at the bottom of the accord reads “President of the Unites States,” misspelling “United.” The error was first spotted by Bloomberg News correspondent David Gura and subsequently highlighted in the president’s Truth Social post and a White House communications email. It is unclear how the typo occurred or whether it will be reflected in the Federal Register, the official repository for presidential actions.

The White House announced a voluntary accord on Tuesday, branding artificial intelligence as “Super Intelligence.” President Donald Trump signed the document alongside six tech executives: Dario Amodei (Anthropic), Elon Musk (SpaceX AI), Greg Brockman (OpenAI), Jensen Huang (Nvidia), Mark Zuckerberg (Meta), and Sundar Pichai (Google). The accord commits signatories to four layers of controls, including robust internal monitoring, independent external auditors, board oversight committees, and safeguards against AI misuse in cybersecurity, biosecurity, and chemical threats.

The Guardian identified a typographical error in the president’s signature at the bottom of the accord, where “United” is misspelled as “Unites.” The mistake was first reported by Bloomberg’s David Gura and subsequently highlighted in the president’s Truth Social post and a White House communications email. The source does not explain how the error occurred, and it is unknown whether the misspelling will be reproduced in the Federal Register, the official archive for presidential actions.

來源詳情: theguardian.com ↗

為什麼這很重要

The accord represents a high‑profile, voluntary attempt by the U.S. government and major AI firms to self‑regulate a rapidly advancing technology amid growing public concern. By committing to robust internal controls, independent audits, and board‑level oversight, the signatories aim to pre‑empt stricter legislation while addressing fears about AI misuse in cybersecurity, biosecurity, and other high‑risk domains. The typo, while minor, underscores the rushed nature of the initiative and may affect perceptions of the administration’s attention to detail in a critical policy arena. Moreover, the uncertainty about the document’s formal filing raises questions about the enforceability and transparency of the accord’s provisions.

The accord is a notable policy development because it reflects a rare collaboration between the U.S. executive branch and leading AI firms to address safety concerns without legislative mandates. By establishing internal controls and external audits, the signatories aim to demonstrate responsible development practices and potentially stave off more stringent regulation.

The typo, while seemingly trivial, draws attention to the speed and perhaps the lack of thorough review in the accord’s preparation. In a climate of heightened scrutiny over AI risks, such an oversight could be leveraged by critics to question the seriousness of the administration’s commitment to .

Interactive Mechanism

互動機制:它實際上是如何運作的

以互動方式探索這項發展背後的基礎技術。

Model Parameter Size:8B Parameters
VRAM Required5.5 GBGPU memory footprint
Target HardwareMacBook / Single GPUDeployment tier
Privacy100% Air-GappedLocal device capability
Core takeaway: Small, quantized models (3B–8B) now run directly inside smartphones and laptops with complete data privacy, while mammoth 400B+ models remain the domain of datacenter clusters.
互動式概念檢查+10 Points
AI Ethics Quiz

Why can ethical evaluation not be reduced to one model score?

接下來看什麼

Observers should monitor whether the misspelling appears in the Federal Register, which could signal the accord’s official status. Future developments may include formal codification of the outlined controls into regulations or laws—a step President Trump has previously resisted. Tracking the implementation of the internal and external audit mechanisms by the six companies will also reveal the practical impact of the voluntary framework. Finally, any public response from consumer advocacy groups or congressional committees could shape the next phase of U.S. .

Whether the misspelling appears in the Federal Register will indicate the accord’s formal status and could affect its legal standing.

Implementation of the outlined controls by the six companies will be a key metric for assessing the accord’s effectiveness and may set precedents for industry‑wide self‑regulation.

Potential moves by Congress or consumer advocacy groups to demand formal regulation could test the durability of the voluntary framework.

相關指引和測驗

AI 倫理AI 的未來什麼是人工智慧?測試你所知道的—嘗試免費的人工智慧測驗在我們的詞彙表中尋找人工智慧術語關注AI監管追蹤器
覺得有用嗎?