AI engineering glossary · Evals guardrails AI security
What is data poisoning?
Data poisoning (also called training data poisoning): An attack that corrupts a model by inserting harmful or misleading examples into its training, fine-tuning or retrieval data so that it learns wrong behaviour or a hidden trigger. Because models absorb whatever they are fed, unvetted data sources are a security risk.
Data poisoning: an example
An attacker seeds a public website with false product claims that get scraped into a fine-tuning dataset, making the model repeat them.
Related terms
More Evals guardrails AI security terms
- DeepEval
- Defence in depth
- Denial of wallet
- Direct prompt injection
- Egress filtering
- Elo rating
- Encoding attack
- Eval set
Where you learn it
Data poisoning is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.