AI hacking incidents
Every record is a real attack in which AI was the weapon, the target or the way in, with the sources that confirm it. Filter and sort the table; the address keeps your filters, so a view can be shared.
| Record | Incident | AI role | Evidence | |||
|---|---|---|---|---|---|---|
| AHL-2025-0001 | Sep 2025 | GTG-1002: Claude Code used to run a state espionage campaign | Weapon | 4 | E1 | Not disclosed |
| AHL-2025-0003 | 24 Aug 2025 to 27 Aug 2025 | s1ngularity: hijacked Nx packages made AI CLIs hunt secrets | Weapon | 2 | E1 | Not disclosed |
| AHL-2025-0002 | 2025 | GTG-2002: Claude Code used for data theft and extortion | Weapon | 3 | E1 | Not disclosed |
No records match these filters. Clear a filter or search for a shorter word.
How records are rated
- Severity, 1 to 5
- Scale of the harm: money lost, records, organisations, critical systems. Rubric
- Evidence, E1 or E2
- E1 is confirmed by the victim, a government or the AI vendor in a primary document; E2 rests on one security vendor's analysis of its own data. Methodology
Download
The full registry under CC BY 4.0, rebuilt with every change. Data dictionary.