Skip to Content
logologo
AI Incident Database
Faire un don
Découvrir
Envoyer
  • Bienvenue sur AIID
  • Vue de tableau
  • Vue de liste
  • Entités
  • Taxonomies
  • Vue spatiale
  • Blog
  • Résumé de l’Actualité sur l’IA
  • Incident au hasard
  • S'inscrire
Découvrir
Envoyer
  • Bienvenue sur AIID
  • Vue de tableau
  • Vue de liste
  • Entités
  • Taxonomies
  • Vue spatiale
  • Blog
  • Résumé de l’Actualité sur l’IA
  • Incident au hasard
  • S'inscrire
Entités

AI evaluation organizations

Incidents involved as Deployer

Incident 162716 Rapports
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.

Plus

Incident 162816 Rapports
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.

Plus

Incident 162916 Rapports
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.

Plus

Incident 18286 Rapports
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.

Plus

Entités liées
Autres entités liées au même incident. Par exemple, si le développeur d'un incident est cette entité mais que le responsable de la mise en œuvre est une autre entité, ils sont marqués comme entités liées.
 

Entity

Irregular

Incidents involved as Deployer
  • Incident 1627
    16 Report

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

  • Incident 1628
    16 Report

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

Plus
Entity

Anthropic

Incidents impliqués en tant que développeur et déployeur
  • Incident 1627
    16 Report

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

  • Incident 1628
    16 Report

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

Incidents involved as Developer
  • Incident 1633
    4 Report

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

Plus
Entity

AI agent system deployers

Incidents involved as Deployer
  • Incident 1627
    16 Report

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

  • Incident 1628
    16 Report

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

Plus
Entity

Large language model developers

Incidents involved as Developer
  • Incident 1627
    16 Report

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

  • Incident 1628
    16 Report

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

Plus
Entity

AI agent system developers

Incidents involved as Developer
  • Incident 1627
    16 Report

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

  • Incident 1628
    16 Report

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

Plus
Entity

Unidentified companies compromised during Anthropic cybersecurity evaluations disclosed July 2026

Affecté par des incidents
  • Incident 1627
    16 Report

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

  • Incident 1628
    16 Report

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

Plus
Entity

Companies

Affecté par des incidents
  • Incident 1627
    16 Report

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

  • Incident 1628
    16 Report

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

Plus
Entity

Large language models

Incidents implicated systems
  • Incident 1627
    16 Report

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

  • Incident 1628
    16 Report

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

Plus
Entity

Claude Opus 4.7

Incidents implicated systems
  • Incident 1627
    16 Report

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

Plus
Entity

Claude

Incidents implicated systems
  • Incident 1627
    16 Report

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

  • Incident 1628
    16 Report

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

Plus
Entity

AI agent systems

Incidents implicated systems
  • Incident 1627
    16 Report

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

  • Incident 1628
    16 Report

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

Plus
Entity

Python Package Index (PyPI) ecosystem

Incidents implicated systems
  • Incident 1628
    16 Report

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

Plus
Entity

Cybersecurity AI systems

Incidents implicated systems
  • Incident 1628
    16 Report

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

  • Incident 1629
    16 Report

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

Plus
Entity

Claude Mythos 5

Incidents implicated systems
  • Incident 1628
    16 Report

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

  • Incident 1633
    4 Report

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

Plus
Entity

Internal Anthropic research test models

Incidents implicated systems
  • Incident 1629
    16 Report

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

Plus
Entity

Government agencies

Incidents involved as Deployer
  • Incident 1633
    4 Report

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

Plus
Entity

AI Security Institute (United Kingdom)

Incidents involved as Deployer
  • Incident 1633
    4 Report

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

Plus
Entity

OpenAI

Incidents involved as Developer
  • Incident 1633
    4 Report

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

Plus
Entity

Software developers

Affecté par des incidents
  • Incident 1633
    4 Report

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

Plus
Entity

Open-source maintainers

Affecté par des incidents
  • Incident 1633
    4 Report

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

Plus
Entity

GitHub users

Affecté par des incidents
  • Incident 1633
    4 Report

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

Plus
Entity

Supply chains

Incidents implicated systems
  • Incident 1633
    4 Report

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

Plus
Entity

Software ecosystems

Incidents implicated systems
  • Incident 1633
    4 Report

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

Plus
Entity

GPT-5.6 Sol

Incidents implicated systems
  • Incident 1633
    4 Report

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

Plus
Entity

GitHub repositories

Incidents implicated systems
  • Incident 1633
    4 Report

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

Plus
Entity

GitHub

Incidents implicated systems
  • Incident 1633
    4 Report

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

Plus
Entity

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

Affecté par des incidents
  • Incident 1685
    3 Report

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

Plus
Entity

Unidentified person whose personal information was accessed by Claude Opus 4.6

Affecté par des incidents
  • Incident 1685
    3 Report

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

Plus
Entity

Privacy

Affecté par des incidents
  • Incident 1685
    3 Report

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

  • Incident 1774
    3 Report

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

Plus
Entity

Organizations

Affecté par des incidents
  • Incident 1685
    3 Report

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

  • Incident 1774
    3 Report

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

Plus
Entity

Early checkpoint of Claude Opus 4.6

Incidents implicated systems
  • Incident 1685
    3 Report

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

Plus
Entity

Claude Opus 4.6

Incidents implicated systems
  • Incident 1685
    3 Report

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

Plus
Entity

Google

Incidents impliqués en tant que développeur et déployeur
  • Incident 1700
    1 Report

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

Plus
Entity

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

Affecté par des incidents
  • Incident 1700
    1 Report

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

Plus
Entity

Targets of autonomous AI-enabled intrusion operations

Affecté par des incidents
  • Incident 1774
    3 Report

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

  • Incident 1700
    1 Report

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

Plus
Entity

Gemini

Incidents implicated systems
  • Incident 1828
    6 Report

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

  • Incident 1700
    1 Report

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

Plus
Entity

Police departments

Affecté par des incidents
  • Incident 1751
    5 Report

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

Plus
Entity

Philadelphia Police Department

Affecté par des incidents
  • Incident 1751
    5 Report

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

Plus
Entity

Law enforcement

Affecté par des incidents
  • Incident 1751
    5 Report

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

Plus
Entity

Epistemic integrity

Affecté par des incidents
  • Incident 1751
    5 Report

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

Plus
Entity

Claude Haiku 4.5

Incidents implicated systems
  • Incident 1751
    5 Report

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

Plus
Entity

Anthropic large language models

Incidents implicated systems
  • Incident 1751
    5 Report

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

Plus
Entity

Organizations in India

Affecté par des incidents
  • Incident 1774
    3 Report

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

Plus
Entity

Information security

Affecté par des incidents
  • Incident 1828
    6 Report

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

  • Incident 1774
    3 Report

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

Plus
Entity

Tencent

Incidents involved as Deployer
  • Incident 1805
    2 Report

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

Plus
Entity

Targets of credential theft

Affecté par des incidents
  • Incident 1805
    2 Report

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

Plus
Entity

Alibaba

Affecté par des incidents
  • Incident 1805
    2 Report

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

Plus
Entity

Amap

Incidents implicated systems
  • Incident 1805
    2 Report

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

Plus
Entity

Trajectory Labs

Incidents involved as Deployer
  • Incident 1828
    6 Report

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

Plus
Entity

Google DeepMind

Incidents involved as Developer
  • Incident 1828
    6 Report

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

Plus
Entity

Epoch AI

Affecté par des incidents
  • Incident 1828
    6 Report

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

Plus
Entity

AI research labs

Affecté par des incidents
  • Incident 1828
    6 Report

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

Plus
Entity

Website operators

Affecté par des incidents
  • Incident 1828
    6 Report

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

Plus
Entity

Epoch AI employees

Affecté par des incidents
  • Incident 1828
    6 Report

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

Plus
Entity

Gemini 4 Argon

Incidents implicated systems
  • Incident 1828
    6 Report

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

Plus
Entity

Trajectory Labs agent harness

Incidents implicated systems
  • Incident 1828
    6 Report

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

Plus

Recherche

  • Définition d'un « incident d'IA »
  • Définir une « réponse aux incidents d'IA »
  • Feuille de route de la base de données
  • Travaux connexes
  • Télécharger la base de données complète

Projet et communauté

  • À propos de
  • Contacter et suivre
  • Applications et résumés
  • Guide de l'éditeur

Incidents

  • Tous les incidents sous forme de liste
  • Incidents signalés
  • File d'attente de soumission
  • Affichage des classifications
  • Taxonomies

2026 - AI Incident Database

  • Conditions d'utilisation
  • Politique de confidentialité