Explanation:
Prompt injection attacks involve carefully crafting malicious inputs (prompts) to manipulate the behavior of a generative AI model during its operation phase, often to bypass safety filters or alter its intended output. In contrast, data poisoning happens during the training phase, model inversion aims to extract training data, and denial-of-service aims to make the system unavailable.
Ultimate access to all questions.
Q.20 Generative AI models are susceptible to various attacks. What type of attack aims to manipulate the model's behavior during operation by carefully crafting inputs?
A
Data poisoning attacks.
B
Model inversion attacks.
C
Prompt injection attacks.
D
Denial-of-service attacks.
No comments yet.