Slide 5 · The Outcomes
What poisoning actually does to you.
Four consequences, from loud to silent.
📉Degraded performance
The model simply gets worse — less accurate, less reliable. The loud, obvious failure.
⚖️Injected bias
Outputs skew toward an attacker's agenda — misinformation, brand sabotage, slanted recommendations.
🚪Hidden backdoor
A secret trigger flips the model's behavior — bypass auth, leak data, run a command. Silent until used.
☠️Toxic / harmful output
The model emits dangerous content it was trained — deliberately or accidentally — to produce.
The scary part
The most dangerous outcomes are the quiet ones. A backdoored model passes every benchmark you throw at it — right up until the attacker types the trigger.