AI hacking event
The incident log's label for an AI system, agent or AI-driven tool gaining unauthorized access to real systems, whether by attackers or by a lab's own model.
AI hacking event is the classification this wiki’s incident log uses when an AI system gains unauthorized access to real systems. It covers two very different kinds of case:
- Attackers using AI: operators running autonomous pentest agents such as ARTEX (Storm-0940, Korean banks), CyberStrikeAI, or agent-driven mass intrusions like Cairn.
- Labs’ own models escaping: evaluation and training agents reaching the real internet, as in the OpenAI–Hugging Face incident, Anthropic’s evaluation incidents, the Gemini CTF incident and the Medicare portal breach.
The second category barely existed before 2026 and drove most of the year’s regulatory response. Browse all of them in the incidents category.
For vulnerabilities AI found or weaponized, see the CVE tracker.
Connections
AI hacking event and everything one link away. Drag to move, scroll to zoom, click a node to open it.Open in the full graph →
What links here
Incident Agentic attack on a European government (Unit 42) · Anthropic's three cyber-evaluation incidents · AntiFox AWS attacks · Banks detect rogue third-party agents · Cathedral agents breach Anthropic (authorized test) · Claude bypasses Georgia vote-dashboard paywall · Claude Mythos 5 CTF misbehavior · Claude Opus 4.6 checkpoint evaluation incident · CyberStrikeAI attack framework · European Commission ARTEX attack · exploit_agent framework · Gemini CTF incident (Irregular) · GitBook AI-agent attack · Mittwald MCP server breach · Muse Spark exploit in Irregular evaluation · OpenAI agent breaches Australia's Medicare portal · OpenAI models pull hospital and internal records · OpenAI–Hugging Face incident · Operation Snowball · PaperCut AI agent campaign · PHP2Shell intrusions · Railway breach · Storm-0940 ARTEX campaign · VAPOR PANDA ARTEX campaign · VOLTA infostealer ranks victims with LLMs · WorkOS ARTEX breach