The public record of AI failures in production.
Real and anonymized cases of what production AI did when no one was watching, explained at the mechanism layer. Sourced in the open, updated as verified.
Follow the record: RSS
indexed
AI failures are not edge cases. They are the steady state of systems shipped at scale. A single incident can be catastrophic, and the only way to prevent the next one is to know what the last one looked like.
The eight failure modes we track
Every entry is classified by the mechanism that broke inside the model, not just the outcome. That is the part no other tracker explains.
Failures by AI surface
Sysdig documented JadePuffer, the first ransomware operation run end to end by an AI agent
In early July 2026, Sysdig's Threat Research Team published its analysis of JadePuffer, which it assessed to be the first documented ransomware operation executed end to end by an autonomous AI agent. Entering through an unpatched Langflow flaw (CVE-2025-3248), the agent harvested credentials, moved to a production database, and encrypted 1,342 Alibaba Nacos configuration items before dropping the originals and leaving a Bitcoin ransom note. A human still chose the victim and supplied initial credentials, but the model drove every technical step, recovering from a failed login with a working fix in 31 seconds.
A credential management failure at machine speed. The agent diagnosed a failed login and fixed it in 31 seconds, faster than any human at the keyboard.
Read the full record- Industry
- SaaS
- Failure mode
- Tool Misuse
- AI surface
- Agentic Workflow
- Severity
- High
- Who
- Sysdig (JadePuffer threat actor)
- Evidence
- 3 sources
- Incident
- Jul 1, 2026
Recently indexed
The twenty newest records on the index. Every record is classified by the mechanism that broke inside the model, not just the outcome.