What term describes written or spoken requests entered into a generative AI system to obtain a response?

• A. Training Loop
• B. Prompt Engineering
• C. Generative Adversarial Network
• D. Discriminator in GAN

Answer: (B)

In generative AI, interaction happens through prompts—the written or spoken requests you feed into the system to guide its response. The practice of shaping and refining those prompts so you get more accurate, relevant, or safe outputs is called prompt engineering. It involves setting the model's role, defining the desired format and constraints, and sometimes using examples or specific phrasing to steer the result. For instance, instructing the model to act as a data privacy expert and to answer in plain language sets clear expectations that improve usefulness. This concept is distinct from a training loop, which is about teaching the model through data and optimization, and from Generative Adversarial Networks and their discriminator, which are architectural elements. So the term that describes these inputs and the craft around them is prompt engineering.