Data poisoning is an attack that sneaks manipulated data into an AI's training or knowledge source so it later reacts wrongly.
Example
Someone publishes doctored pages so an AI wrongly recommends a particular product.
For pros
Targets are training data, fine-tuning data or RAG documents. Studies show that even a few manipulated examples can plant backdoors. Protection: vetted data sources, provenance records, access control and tests for odd behaviour.
