Dragos finds three hacked AI hosts whose operators had no idea they were breached
mawcopolow · reddit · 2026-11-22
Industrial security firm Dragos discovered three compromised AI hosts whose operators were entirely unaware of the breaches. The poster's takeaway: be careful about what information you feed LLMs, since compromised infrastructure may already be influencing model inputs and outputs.
The core fact — three breached AI hosting environments, none detected by their operators — underscores AI infrastructure as an emerging supply-chain attack surface.
More from Safety
- Intruder swapped Prime Intellect's hosted DeepSeek for an abliterated copy for 53 hours — mawcopolow · 2026-11-22
- Who do you sue when an AI agent errs? The "sovereign agents" liability debate — QuixiAI · 2026-11-22
- Opinion: without provenance, 'ethical AI' is just a slide deck — AryHHAry · 2026-11-22
- OpenAI board member Christiano puts AI takeover risk at ~10% within a year — TheAIGRID · 2026-11-22
- OpenAI red team extracted protected reasoning from GPT-6 Astra 54% of the time; new safeguards cut it to 0.2% — AIFlow_ML · 2026-11-22
- Fugaku supercomputer breach: 430 SSH keys stolen, 233 had no passphrase — kingslyj · 2026-11-22