Generative Models and Prompting

This unit explores a range of generative AI systems (audio, image, code, and face-swap) and how prompt design influences their outputs, strengths, limitations, and futures. Students will practice model-specific prompt engineering, assess biases and cultural sensitivities, and compile documented use cases and recommendations to prepare for advanced application and capstone work.

Learning Objectives

• Analyze differences in prompt sensitivity, output modalities, and common failure modes across multiple generative systems (e.g., MuseNet, BigGAN, Codex, Midjourney, RunwayML, DeepFaceLab)
• Apply model-specific prompt-engineering techniques to iteratively create and refine prompts for at least three generative tools and measure output improvements using defined quality criteria (e.g., fidelity, relevance, style accuracy)
• Evaluate ethical concerns, biases, and cultural sensitivity risks present in generated outputs and propose concrete mitigation strategies aligned with responsible AI best practices
• Demonstrate professional documentation skills by producing a comparative portfolio that summarizes use cases, sample prompts, limitations, and recommendations for future development and responsible deployment