Glossary: Application and Credential Attacks, and AI Risks
poisoning
Also written: data poisoning, training data poisoning
Putting bad data into what an AI learns from, such as training data, a knowledge source or user feedback, so the model learns the wrong lessons. The effect lasts until the model is retrained on clean data.
Where the book teaches it
Related terms
backdoorevasionmodel manipulationprovenancetraining dataknowledge sourcehallucination