PulseAugur
EN
LIVE 13:28:39
한국어(KO) 트럼프 행정부가 앤트로픽(Anthropic)이 사전보고한 것보다 더 많은 기관에 최신 모델 'Mythos' 접근권을 준 사실과 아마존의 취약점 보고를 근거로 수출통제 검토를 압박해 앤트로픽이 모델을 내리게 했다. 정부는 민감기술 보호 실패를 문제삼았고, 업계는 선례·혁신 저해를 우려한다

US government restricts Anthropic AI models over security concerns

The U.S. government has imposed export controls on Anthropic's AI models, including 'Mythos' and 'Fable 5', leading the company to withdraw them globally. This action was prompted by concerns from Amazon regarding security risks and Anthropic's alleged failure to adequately address potential vulnerabilities, such as 'jailbreaking'. The move has sparked debate within the industry about the balance between AI innovation and national security, with some experts questioning the broad impact of such controls. AI

IMPACT Sets a precedent for government intervention in AI model distribution, potentially impacting future innovation and global access.

RANK_REASON Government regulatory action impacting a major AI company's product release.

Read on Mastodon — sigmoid.social →

AI-generated summary · Google Gemini · from 8 sources. How we write summaries →

US government restricts Anthropic AI models over security concerns

COVERAGE [8]

  1. Mastodon — sigmoid.social TIER_1 한국어(KO) · [email protected] ·

    The Trump administration pressured Anthropic to take down its latest model 'Mythos' based on the fact that it granted access to more agencies than Anthropic had pre-reported, and Amazon's vulnerability report. The government cited failures in protecting sensitive technology, while the industry expressed concerns about precedents and hindering innovation.

    트럼프 행정부가 앤트로픽(Anthropic)이 사전보고한 것보다 더 많은 기관에 최신 모델 'Mythos' 접근권을 준 사실과 아마존의 취약점 보고를 근거로 수출통제 검토를 압박해 앤트로픽이 모델을 내리게 했다. 정부는 민감기술 보호 실패를 문제삼았고, 업계는 선례·혁신 저해를 우려한다. https://www. washingtonpost.com/technology/ 2026/06/15/how-anthropic-lost-white-houses-trust-then-its-flagship-product…

  2. Mastodon — sigmoid.social TIER_1 한국어(KO) · [email protected] ·

    ITmedia NEWS (@itmedia_news) reports that Amazon expressed concerns about Anthropic's latest AI models/systems, which appear to be linked to a US administration order to halt operations. This suggests a debate on trust and safety in AI, as well as regulatory issues among major AI companies. https:// x.c

    ITmedia NEWS (@itmedia_news) 아마존이 Anthropic의 최신 AI 모델/시스템에 대해 우려를 전달했으며, 이후 미국 행정부의 중단 명령과 연관된 것으로 보인다는 보도입니다. 주요 AI 기업 간의 신뢰·안전성 논쟁과 규제 이슈를 시사합니다. https:// x.com/itmedia_news/status/2066 090792905810091 # amazon # anthropic # ai # safety # regulation

  3. Mastodon — mastodon.social TIER_1 English(EN) · Wesearchpress ·

    US government imposes export control on Anthropic's Fable 5 AI model, citing potential dangers, amid dispute with the company # ai # regulation https:// wesearc

    US government imposes export control on Anthropic's Fable 5 AI model, citing potential dangers, amid dispute with the company # ai # regulation https:// wesearch.press/s/tyler-cowen-a -dangerous-turn-in-ai-regulation-8502efde?utm_source=social&utm_medium=auto&utm_campaign=mastodo…

  4. Mastodon — mastodon.social TIER_1 日本語(JA) · [email protected] ·

    Amazon CEO's concerns about Anthropic's latest AI indicate that AI risk management has entered a phase involving not only development companies but also cloud providers. Large-scale AI does not operate solely as a model but is connected to computing resources, data storage, and services for governments and businesses. Therefore, advanced AI poses risks such as fraud, cyberattacks, and confidential information...

    "アマゾンCEOがアンソロピックの最新AIに懸念を伝えたことは、AIのリスク管理が開発企業だけでなく、クラウド事業者にも関わる段階に入ったことを示しています。" "大規模AIは、モデル単体で動くのではなく、計算資源、データ保管、政府や企業向けサービスと結び付いています。そのため、高度なAIが詐欺、サイバー攻撃、機密情報の探索などに使われるおそれがあれば、提供方法そのものを管理する必要があります。" "一方で、外国人アクセスの全面停止や世界的停止は、判断基準が不明確なまま進めば、市場排除にも見えかねません。必要なのは、危険性の説明、利用者審査、利用履歴の監…

  5. Mastodon — mastodon.social TIER_1 日本語(JA) · [email protected] ·

    Some experts supporting export controls on advanced AI models find the Trump administration's move baffling because it affects not only adversaries but also allies. "This is not well thought out," said Jimmy Goodrich, a senior fellow at the University of California's Institute on Global Conflict and Cooperation. "Canada, employed by Anthropic

    "高度なAIモデルに対する輸出規制を支持する一部の専門家は、トランプ政権のこの措置が敵対国だけでなく同盟国にも影響を与えるため、不可解だと感じている。" "「これはよく考えられていない」と、カリフォルニア大学グローバル紛争協力研究所の上級研究員、ジミー・グッドリッチ氏は述べた。「Anthropic社に雇用されているカナダ人やイギリス人さえも研究開発に従事することを禁じている」。" "この命令は、トランプ政権当局者とAnthropic社との間の以前の紛争が、米国政府の一部で緩和の兆しを見せ始めた矢先に出された。" (Google翻訳) https://ww…

  6. Mastodon — mastodon.social TIER_1 日本語(JA) · [email protected] ·

    Anthropic stated in a blog post that the US government's regulation was carried out in the form of export controls. White House advisor David Sacks said in a social media post on Saturday that authorities "reluctantly" issued export controls because Anthropic CEO Dario Amodei "refused" to "fix jailbreaks or remove models."

    "Anthropic社はブログ記事で、米政府による規制は輸出管理という形で行われたと述べた。" "ホワイトハウス顧問のデビッド・サックス氏は土曜日のソーシャルメディアへの投稿で、Anthropic社のダリオ・アモデイCEOが「脱獄の修正やモデルの削除」を「拒否」したため、当局は「不本意ながら」輸出管理を発令したと述べた。" "「政権の現在の希望は、Anthropic社が安全性の問題を解決し、輸出規制が解除され、『Fable』が一般発売に戻ることだ」と、トランプ政権の科学技術諮問委員会の共同議長で、以前はホワイトハウスのAI担当責任者だったサックス氏は書い…

  7. Mastodon — mastodon.social TIER_1 日本語(JA) · [email protected] ·

    Anthropic had previously warned about the hacking capabilities of its Mythos model and refrained from making it public, but earlier this week, it released a public version named Fable, claiming to have implemented cybersecurity measures. The release ended on Friday. In a blog post, Anthropic stated that the US government, using this model for cybersecurity

    "Anthropicは、以前から自社のMythosモデルのハッキング能力について警告し、一般公開を控えていたが、今週初め、サイバーセキュリティ対策を施したと称するFableという名称の公開版をリリースした。" "公開は金曜日に終了した。Anthropicはブログ記事で、米国政府が、このモデルを使ってサイバーセキュリティの脆弱性を発見するための対策を回避する、いわゆる「脱獄」の方法が存在すると考えていると伝えてきたと述べた。" "米国内外を問わず、外国人が最新モデルであるFable 5とMythos 5の両方を使用することを禁止するようAnthropicに…

  8. Mastodon — mastodon.social TIER_1 日本語(JA) · [email protected] ·

    Amazon CEO Andy Jassy was one of the tech leaders who expressed concerns about the security risks of Anthropic's cutting-edge AI models to Trump administration officials this week, sources familiar with the matter told Reuters. Jassy's involvement comes amid the Trump administration's national security order...

    "Amazonのアンディ・ジャシーCEOは、今週、トランプ政権高官に対し、Anthropic社の最先端AIモデルにおけるセキュリティリスクについて懸念を表明したテクノロジー業界のリーダーの一人だったと、事情に詳しい関係者がロイター通信に語った。" "ジャシー氏の関与は、ドナルド・トランプ大統領政権の国家安全保障命令を受け、Anthropic社が金曜日に最新モデルを世界中で停止するという異例の措置をとった背景を明らかにするものだ。" "Amazonは、Anthropic社のモデルについて政府関係者と協議したかどうかについては確認しなかった。" (Googl…