Module Check: Generative AI
You have traveled through the full landscape of generative AI: from the core idea of creating new content rather than classifying existing content, through the specific mechanisms of text, image, audio, and video generation, into the unified world of multimodal systems, and out through the practical territory of using these tools well. This lesson brings it all together. Start with a review of the key terms, test your understanding across the full module, and finish with a synthesis activity that shows what you can now do with these ideas.
Key Terms Review
Flashcards — click each card to reveal the answer
Module Quiz
A system classifies every submitted resume as 'strong candidate' or 'weak candidate.' A second system generates a personalized cover letter for any job description the user inputs. Which is generative AI and what is the key difference?
During text generation, the model assigns a probability of 72% to the word 'Paris,' 15% to 'London,' and smaller probabilities to other words. The temperature is set very high. What is most likely to happen?
Why do current AI image generators typically produce scrambled or incorrect text when asked to include readable words in an image?
What makes temporal consistency the central challenge of AI video generation?
A student asks an AI: 'What were the top news stories from two weeks ago?' The AI confidently lists five stories with specific details. What should the student do and why?
A multimodal AI is shown a photo of a handwritten math problem and asked to solve it step by step. Which capability makes this possible?
Generative AI creates new content by learning statistical patterns in human-made text, images, audio, and video — and applying those patterns to produce something that did not exist before. It can produce remarkable outputs at remarkable speed, and it fails in predictable, structural ways. Understanding both sides — the power and the limits — is what makes you a skilled, responsible participant in the world generative AI is shaping.
Synthesis: Generative AI in the Wild
- Step 1: Choose one real domain where generative AI is already being used or is likely to be used soon. Options: journalism, medicine, music, film production, education, scientific research, legal services, video game development, fashion design, or a domain of your own choosing.
- Step 2: Describe three specific ways generative AI is being or could be applied in that domain. For each application, name the modality involved (text, image, audio, video, or multimodal).
- Step 3: For each of the three applications, identify one genuine benefit and one genuine risk or limit. Be specific — reference the failure modes from Lesson 7 or the ethical issues from Lessons 4, 5, and 8.
- Step 4: Write a one-paragraph policy recommendation — aimed at a school board, a hospital administrator, or a local government, depending on the domain you chose — that addresses how your domain should approach generative AI responsibly. Your recommendation should acknowledge both the value and the risks.
- Step 5: Review your policy recommendation with this question in mind: does it account for the students, patients, or community members who would be most affected if the AI failed? Revise if needed, then write a final sentence explaining why accounting for vulnerable users is an essential part of responsible AI policy.