Data poisoning
AIGP glossary · Last reviewed: · By Victor Humenhuk (AIGP certified)
Data poisoning - Corrupting the training data so the model learns wrong or malicious behaviour.
In the AIGP body of knowledge, Data poisoning comes up under Module 3: Governance & Risk Management · BoK III.A; Module 8: AI Governance Vocabulary.
Data poisoning in context
- The [[Security risk]] card carries the most testable vocabulary → [[Adversarial attack]]s, [[Hallucination]]s, [[Deepfake]]s and [[Data poisoning]]. (The four AI risk categories)
- ==Intent is the dividing line==: disinformation is deliberate, misinformation is not - and data poisoning is an attacker corrupting training data, unlike natural data drift. (Risks, security and harms)
Where Data poisoning is covered in the AIGP study notes
Related terms
- Operational risk
- Legal risk
- Security risk
- Privacy risk
- Adversarial attack
- Deepfake
- Hallucination
- Data persistence
- Data spillover
- Data leak
Test yourself on Data poisoning
Recognising a definition is not the same as applying it in an exam scenario. Work through the free AIGP practice questions, or read the full AIGP study guide - every study note is free.