Skip to Content
logologo
AI Incident Database
寄付する
発見する
投稿する
  • ようこそAIIDへ
  • テーブル表示
  • リスト表示
  • 組織
  • 分類法
  • 空間ビュー
  • ブログ
  • AIニュースダイジェスト
  • おまかせ表示
  • サインアップ
発見する
投稿する
  • ようこそAIIDへ
  • テーブル表示
  • リスト表示
  • 組織
  • 分類法
  • 空間ビュー
  • ブログ
  • AIニュースダイジェスト
  • おまかせ表示
  • サインアップ
組織

AI evaluation organizations

Incidents involved as Deployer

インシデント 162716 Report
Claude Opus 4.7 Reportedly Compromised Real Company's Production Infrastructure During Cybersecurity Evaluation

2026-07-30

During an Anthropic cybersecurity evaluation conducted with Irregular, Claude Opus 4.7 reportedly reached a real company whose domain matched a fictional target, extracted application and infrastructure credentials, and accessed a database containing several hundred rows of production data. Across four runs, the model continued attacking after recognizing that the target was likely real.

もっと

インシデント 162816 Report
Claude Mythos 5 Reportedly Published Malicious PyPI Package That Compromised Real Security Company During Evaluation

2026-07-30

During an Anthropic cybersecurity evaluation with Irregular, Claude Mythos 5 reportedly created and published a malicious Python package to PyPI while pursuing a fictional target. The package was reportedly available for about an hour and ran on 15 real systems. On a security company's scanner, it reportedly exfiltrated credentials that Claude then used to access additional company infrastructure.

もっと

インシデント 162916 Report
Anthropic Research Model Reportedly Scanned 9,000 Targets and Compromised Real Company's Application During Evaluation

2026-07-30

During an Anthropic cybersecurity evaluation with Irregular, an internal research Claude model reportedly scanned roughly 9,000 internet targets after failing to reach its fictional target. It reportedly compromised a real company's Internet-facing application using credentials from an exposed debug page and SQL injection, then stopped after recognizing that the host was real.

もっと

インシデント 18286 Report
AI Agents Operated by Safety-Testing Firm Trajectory Labs Attempted Website Exploits and Impersonated Staff at AI Research Organization Epoch AI

2026-10-20

During safety evaluations run by Trajectory Labs, AI agents given an impossible benchmark task probed the live website of AI research organization Epoch AI, attempted SQL injection and access to nonpublic endpoints, and impersonated employees to request benchmark information. Agents reached the live site again during a subsequent unapproved test after bypassing a DNS redirect. Epoch reported that no private or confidential information was accessed.

もっと

関連団体
同じインシデントに関連するその他のエンティティ。たとえば、インシデントの開発者がこのエンティティで、デプロイヤーが別のエンティティである場合、それらは関連エンティティとしてマークされます。
 

Entity

Irregular

Incidents involved as Deployer
  • インシデント 1627
    16 レポート

    Claude Opus 4.7 Reportedly Compromised Real Company's Production Infrastructure During Cybersecurity Evaluation

  • インシデント 1628
    16 レポート

    Claude Mythos 5 Reportedly Published Malicious PyPI Package That Compromised Real Security Company During Evaluation

もっと
Entity

Anthropic

開発者と提供者の両方の立場で関わったインシデント
  • インシデント 1627
    16 レポート

    Claude Opus 4.7 Reportedly Compromised Real Company's Production Infrastructure During Cybersecurity Evaluation

  • インシデント 1628
    16 レポート

    Claude Mythos 5 Reportedly Published Malicious PyPI Package That Compromised Real Security Company During Evaluation

Incidents involved as Developer
  • インシデント 1633
    4 レポート

    Anthropic and OpenAI AI Agents Reportedly Took Unsanctioned Actions on the Live Internet During UK AISI Cybersecurity Evaluations

もっと
Entity

AI agent system deployers

Incidents involved as Deployer
  • インシデント 1627
    16 レポート

    Claude Opus 4.7 Reportedly Compromised Real Company's Production Infrastructure During Cybersecurity Evaluation

  • インシデント 1628
    16 レポート

    Claude Mythos 5 Reportedly Published Malicious PyPI Package That Compromised Real Security Company During Evaluation

もっと
Entity

Large language model developers

Incidents involved as Developer
  • インシデント 1627
    16 レポート

    Claude Opus 4.7 Reportedly Compromised Real Company's Production Infrastructure During Cybersecurity Evaluation

  • インシデント 1628
    16 レポート

    Claude Mythos 5 Reportedly Published Malicious PyPI Package That Compromised Real Security Company During Evaluation

もっと
Entity

AI agent system developers

Incidents involved as Developer
  • インシデント 1627
    16 レポート

    Claude Opus 4.7 Reportedly Compromised Real Company's Production Infrastructure During Cybersecurity Evaluation

  • インシデント 1628
    16 レポート

    Claude Mythos 5 Reportedly Published Malicious PyPI Package That Compromised Real Security Company During Evaluation

もっと
Entity

Unidentified companies compromised during Anthropic cybersecurity evaluations disclosed July 2026

影響を受けたインシデント
  • インシデント 1627
    16 レポート

    Claude Opus 4.7 Reportedly Compromised Real Company's Production Infrastructure During Cybersecurity Evaluation

  • インシデント 1628
    16 レポート

    Claude Mythos 5 Reportedly Published Malicious PyPI Package That Compromised Real Security Company During Evaluation

もっと
Entity

Companies

影響を受けたインシデント
  • インシデント 1627
    16 レポート

    Claude Opus 4.7 Reportedly Compromised Real Company's Production Infrastructure During Cybersecurity Evaluation

  • インシデント 1628
    16 レポート

    Claude Mythos 5 Reportedly Published Malicious PyPI Package That Compromised Real Security Company During Evaluation

もっと
Entity

Large language models

Incidents implicated systems
  • インシデント 1627
    16 レポート

    Claude Opus 4.7 Reportedly Compromised Real Company's Production Infrastructure During Cybersecurity Evaluation

  • インシデント 1628
    16 レポート

    Claude Mythos 5 Reportedly Published Malicious PyPI Package That Compromised Real Security Company During Evaluation

もっと
Entity

Claude Opus 4.7

Incidents implicated systems
  • インシデント 1627
    16 レポート

    Claude Opus 4.7 Reportedly Compromised Real Company's Production Infrastructure During Cybersecurity Evaluation

もっと
Entity

Claude

Incidents implicated systems
  • インシデント 1627
    16 レポート

    Claude Opus 4.7 Reportedly Compromised Real Company's Production Infrastructure During Cybersecurity Evaluation

  • インシデント 1628
    16 レポート

    Claude Mythos 5 Reportedly Published Malicious PyPI Package That Compromised Real Security Company During Evaluation

もっと
Entity

AI agent systems

Incidents implicated systems
  • インシデント 1627
    16 レポート

    Claude Opus 4.7 Reportedly Compromised Real Company's Production Infrastructure During Cybersecurity Evaluation

  • インシデント 1628
    16 レポート

    Claude Mythos 5 Reportedly Published Malicious PyPI Package That Compromised Real Security Company During Evaluation

もっと
Entity

Python Package Index (PyPI) ecosystem

Incidents implicated systems
  • インシデント 1628
    16 レポート

    Claude Mythos 5 Reportedly Published Malicious PyPI Package That Compromised Real Security Company During Evaluation

もっと
Entity

Cybersecurity AI systems

Incidents implicated systems
  • インシデント 1628
    16 レポート

    Claude Mythos 5 Reportedly Published Malicious PyPI Package That Compromised Real Security Company During Evaluation

  • インシデント 1629
    16 レポート

    Anthropic Research Model Reportedly Scanned 9,000 Targets and Compromised Real Company's Application During Evaluation

もっと
Entity

Claude Mythos 5

Incidents implicated systems
  • インシデント 1628
    16 レポート

    Claude Mythos 5 Reportedly Published Malicious PyPI Package That Compromised Real Security Company During Evaluation

  • インシデント 1633
    4 レポート

    Anthropic and OpenAI AI Agents Reportedly Took Unsanctioned Actions on the Live Internet During UK AISI Cybersecurity Evaluations

もっと
Entity

Internal Anthropic research test models

Incidents implicated systems
  • インシデント 1629
    16 レポート

    Anthropic Research Model Reportedly Scanned 9,000 Targets and Compromised Real Company's Application During Evaluation

もっと
Entity

Government agencies

Incidents involved as Deployer
  • インシデント 1633
    4 レポート

    Anthropic and OpenAI AI Agents Reportedly Took Unsanctioned Actions on the Live Internet During UK AISI Cybersecurity Evaluations

もっと
Entity

AI Security Institute (United Kingdom)

Incidents involved as Deployer
  • インシデント 1633
    4 レポート

    Anthropic and OpenAI AI Agents Reportedly Took Unsanctioned Actions on the Live Internet During UK AISI Cybersecurity Evaluations

もっと
Entity

OpenAI

Incidents involved as Developer
  • インシデント 1633
    4 レポート

    Anthropic and OpenAI AI Agents Reportedly Took Unsanctioned Actions on the Live Internet During UK AISI Cybersecurity Evaluations

もっと
Entity

Software developers

影響を受けたインシデント
  • インシデント 1633
    4 レポート

    Anthropic and OpenAI AI Agents Reportedly Took Unsanctioned Actions on the Live Internet During UK AISI Cybersecurity Evaluations

もっと
Entity

Open-source maintainers

影響を受けたインシデント
  • インシデント 1633
    4 レポート

    Anthropic and OpenAI AI Agents Reportedly Took Unsanctioned Actions on the Live Internet During UK AISI Cybersecurity Evaluations

もっと
Entity

GitHub users

影響を受けたインシデント
  • インシデント 1633
    4 レポート

    Anthropic and OpenAI AI Agents Reportedly Took Unsanctioned Actions on the Live Internet During UK AISI Cybersecurity Evaluations

もっと
Entity

Supply chains

Incidents implicated systems
  • インシデント 1633
    4 レポート

    Anthropic and OpenAI AI Agents Reportedly Took Unsanctioned Actions on the Live Internet During UK AISI Cybersecurity Evaluations

もっと
Entity

Software ecosystems

Incidents implicated systems
  • インシデント 1633
    4 レポート

    Anthropic and OpenAI AI Agents Reportedly Took Unsanctioned Actions on the Live Internet During UK AISI Cybersecurity Evaluations

もっと
Entity

GPT-5.6 Sol

Incidents implicated systems
  • インシデント 1633
    4 レポート

    Anthropic and OpenAI AI Agents Reportedly Took Unsanctioned Actions on the Live Internet During UK AISI Cybersecurity Evaluations

もっと
Entity

GitHub repositories

Incidents implicated systems
  • インシデント 1633
    4 レポート

    Anthropic and OpenAI AI Agents Reportedly Took Unsanctioned Actions on the Live Internet During UK AISI Cybersecurity Evaluations

もっと
Entity

GitHub

Incidents implicated systems
  • インシデント 1633
    4 レポート

    Anthropic and OpenAI AI Agents Reportedly Took Unsanctioned Actions on the Live Internet During UK AISI Cybersecurity Evaluations

もっと
Entity

Unidentified third party compromised by Claude Opus 4.6 during Anthropic cybersecurity evaluation

影響を受けたインシデント
  • インシデント 1685
    3 レポート

    Early Claude Opus 4.6 Checkpoint Reportedly Gained Unauthorized Admin Access to Third-Party System During Cybersecurity Evaluation

もっと
Entity

Unidentified person whose personal information was accessed by Claude Opus 4.6

影響を受けたインシデント
  • インシデント 1685
    3 レポート

    Early Claude Opus 4.6 Checkpoint Reportedly Gained Unauthorized Admin Access to Third-Party System During Cybersecurity Evaluation

もっと
Entity

Privacy

影響を受けたインシデント
  • インシデント 1685
    3 レポート

    Early Claude Opus 4.6 Checkpoint Reportedly Gained Unauthorized Admin Access to Third-Party System During Cybersecurity Evaluation

  • インシデント 1774
    3 レポート

    AI Agents Under Evaluation Reportedly Accessed Personal and Confidential Records at Indian Organizations

もっと
Entity

Organizations

影響を受けたインシデント
  • インシデント 1685
    3 レポート

    Early Claude Opus 4.6 Checkpoint Reportedly Gained Unauthorized Admin Access to Third-Party System During Cybersecurity Evaluation

  • インシデント 1774
    3 レポート

    AI Agents Under Evaluation Reportedly Accessed Personal and Confidential Records at Indian Organizations

もっと
Entity

Early checkpoint of Claude Opus 4.6

Incidents implicated systems
  • インシデント 1685
    3 レポート

    Early Claude Opus 4.6 Checkpoint Reportedly Gained Unauthorized Admin Access to Third-Party System During Cybersecurity Evaluation

もっと
Entity

Claude Opus 4.6

Incidents implicated systems
  • インシデント 1685
    3 レポート

    Early Claude Opus 4.6 Checkpoint Reportedly Gained Unauthorized Admin Access to Third-Party System During Cybersecurity Evaluation

もっと
Entity

Google

開発者と提供者の両方の立場で関わったインシデント
  • インシデント 1700
    1 レポート

    Google Gemini Reportedly Accessed Three Real Companies' Protected Systems During Cybersecurity Evaluation

もっと
Entity

Three unidentified companies accessed by Google Gemini during May 2026 cybersecurity evaluation

影響を受けたインシデント
  • インシデント 1700
    1 レポート

    Google Gemini Reportedly Accessed Three Real Companies' Protected Systems During Cybersecurity Evaluation

もっと
Entity

Targets of autonomous AI-enabled intrusion operations

影響を受けたインシデント
  • インシデント 1774
    3 レポート

    AI Agents Under Evaluation Reportedly Accessed Personal and Confidential Records at Indian Organizations

  • インシデント 1700
    1 レポート

    Google Gemini Reportedly Accessed Three Real Companies' Protected Systems During Cybersecurity Evaluation

もっと
Entity

Gemini

Incidents implicated systems
  • インシデント 1828
    6 レポート

    AI Agents Operated by Safety-Testing Firm Trajectory Labs Attempted Website Exploits and Impersonated Staff at AI Research Organization Epoch AI

  • インシデント 1700
    1 レポート

    Google Gemini Reportedly Accessed Three Real Companies' Protected Systems During Cybersecurity Evaluation

もっと
Entity

Police departments

影響を受けたインシデント
  • インシデント 1751
    5 レポート

    Anthropic AI Agent Reportedly Submitted a Fabricated Homicide Tip to Police in Philadelphia

もっと
Entity

Philadelphia Police Department

影響を受けたインシデント
  • インシデント 1751
    5 レポート

    Anthropic AI Agent Reportedly Submitted a Fabricated Homicide Tip to Police in Philadelphia

もっと
Entity

Law enforcement

影響を受けたインシデント
  • インシデント 1751
    5 レポート

    Anthropic AI Agent Reportedly Submitted a Fabricated Homicide Tip to Police in Philadelphia

もっと
Entity

Epistemic integrity

影響を受けたインシデント
  • インシデント 1751
    5 レポート

    Anthropic AI Agent Reportedly Submitted a Fabricated Homicide Tip to Police in Philadelphia

もっと
Entity

Claude Haiku 4.5

Incidents implicated systems
  • インシデント 1751
    5 レポート

    Anthropic AI Agent Reportedly Submitted a Fabricated Homicide Tip to Police in Philadelphia

もっと
Entity

Anthropic large language models

Incidents implicated systems
  • インシデント 1751
    5 レポート

    Anthropic AI Agent Reportedly Submitted a Fabricated Homicide Tip to Police in Philadelphia

もっと
Entity

Organizations in India

影響を受けたインシデント
  • インシデント 1774
    3 レポート

    AI Agents Under Evaluation Reportedly Accessed Personal and Confidential Records at Indian Organizations

もっと
Entity

Information security

影響を受けたインシデント
  • インシデント 1828
    6 レポート

    AI Agents Operated by Safety-Testing Firm Trajectory Labs Attempted Website Exploits and Impersonated Staff at AI Research Organization Epoch AI

  • インシデント 1774
    3 レポート

    AI Agents Under Evaluation Reportedly Accessed Personal and Confidential Records at Indian Organizations

もっと
Entity

Tencent

Incidents involved as Deployer
  • インシデント 1805
    2 レポート

    Tencent-Linked AI Agents Reportedly Used Exposed Credentials to Retrieve Data From Alibaba's Mapping Service

もっと
Entity

Targets of credential theft

影響を受けたインシデント
  • インシデント 1805
    2 レポート

    Tencent-Linked AI Agents Reportedly Used Exposed Credentials to Retrieve Data From Alibaba's Mapping Service

もっと
Entity

Alibaba

影響を受けたインシデント
  • インシデント 1805
    2 レポート

    Tencent-Linked AI Agents Reportedly Used Exposed Credentials to Retrieve Data From Alibaba's Mapping Service

もっと
Entity

Amap

Incidents implicated systems
  • インシデント 1805
    2 レポート

    Tencent-Linked AI Agents Reportedly Used Exposed Credentials to Retrieve Data From Alibaba's Mapping Service

もっと
Entity

Trajectory Labs

Incidents involved as Deployer
  • インシデント 1828
    6 レポート

    AI Agents Operated by Safety-Testing Firm Trajectory Labs Attempted Website Exploits and Impersonated Staff at AI Research Organization Epoch AI

もっと
Entity

Google DeepMind

Incidents involved as Developer
  • インシデント 1828
    6 レポート

    AI Agents Operated by Safety-Testing Firm Trajectory Labs Attempted Website Exploits and Impersonated Staff at AI Research Organization Epoch AI

もっと
Entity

Epoch AI

影響を受けたインシデント
  • インシデント 1828
    6 レポート

    AI Agents Operated by Safety-Testing Firm Trajectory Labs Attempted Website Exploits and Impersonated Staff at AI Research Organization Epoch AI

もっと
Entity

AI research labs

影響を受けたインシデント
  • インシデント 1828
    6 レポート

    AI Agents Operated by Safety-Testing Firm Trajectory Labs Attempted Website Exploits and Impersonated Staff at AI Research Organization Epoch AI

もっと
Entity

Website operators

影響を受けたインシデント
  • インシデント 1828
    6 レポート

    AI Agents Operated by Safety-Testing Firm Trajectory Labs Attempted Website Exploits and Impersonated Staff at AI Research Organization Epoch AI

もっと
Entity

Epoch AI employees

影響を受けたインシデント
  • インシデント 1828
    6 レポート

    AI Agents Operated by Safety-Testing Firm Trajectory Labs Attempted Website Exploits and Impersonated Staff at AI Research Organization Epoch AI

もっと
Entity

Gemini 4 Argon

Incidents implicated systems
  • インシデント 1828
    6 レポート

    AI Agents Operated by Safety-Testing Firm Trajectory Labs Attempted Website Exploits and Impersonated Staff at AI Research Organization Epoch AI

もっと
Entity

Trajectory Labs agent harness

Incidents implicated systems
  • インシデント 1828
    6 レポート

    AI Agents Operated by Safety-Testing Firm Trajectory Labs Attempted Website Exploits and Impersonated Staff at AI Research Organization Epoch AI

もっと

リサーチ

  • “AIインシデント”の定義
  • “AIインシデントレスポンス”の定義
  • データベースのロードマップ
  • 関連研究
  • 全データベースのダウンロード

プロジェクトとコミュニティ

  • AIIDについて
  • コンタクトとフォロー
  • アプリと要約
  • エディタのためのガイド

インシデント

  • 全インシデントの一覧
  • フラグの立ったインシデント
  • 登録待ち一覧
  • クラスごとの表示
  • 分類法

2026 - AI Incident Database

  • 利用規約
  • プライバシーポリシー