Generative AI Tools and Prompting

This unit examines a range of generative AI models (audio, image, text/code, and deepfake tools) and how prompt design influences their outputs, biases, and real-world use cases. Students will compare model strengths and limitations, practice model-specific prompting, and assess ethical and cultural considerations to prepare for applied projects in the final unit.

Learning Objectives

• Analyze how prompts influence output characteristics, bias, and cultural sensitivity across multiple generative models (e.g., MuseNet, BigGAN, Codex, RunwayML, DeepArt, Midjourney, Google Bard, DeepFaceLab).
• Apply model-specific prompting techniques to produce targeted outputs in at least three different tool categories (text/code, image, audio/visual), documenting adjustments and observed results.
• Evaluate ethical issues, limitations, and potential harms in model outputs and propose concrete mitigation strategies and improvements for culturally sensitive or biased content.
• Demonstrate the ability to iteratively create, test, and refine sample prompts for selected tools and summarize findings, including recommendations for future development and deployment.