12 · Check what you learned

1. How does generative AI mainly create a new image, song, or paragraph?

It imagines it from its own feelingsIt learns patterns from huge amounts of human-made work, then produces something new that fitsIt copies one existing file and renames itIt looks the answer up in a single book

2. Because AI is built from patterns, what does it tend to produce — and what stays a human job?

It produces genius-level originality; humans just watchIt produces what's *typical*; taste and judgment about what's *good* stay humanIt produces perfect facts every time; humans do nothingIt produces only random noise; humans fix all of it

3. What's the difference between a lazy creative prompt and a strong one?

A strong prompt is just longerA strong prompt directs the AI — subject, tone/style, purpose, restrictions — and reflects your point of viewA strong prompt uses fancier words the AI prefersThere's no difference; results are random

4. What's the healthiest way to think about an AI creative tool?

A replacement for your own ideas and effortA collaborator you direct — it assists your vision, you stay the authorA machine that should do the whole project so you don't have toSomething to avoid completely

5. Why do many artists object to AI being trained on their work?

Artists dislike all new technologyTheir real, human-made work was used to train the model — often without being asked, paid, or creditedThe AI pays them too muchThere's no real reason; it's imaginary

6. Under U.S. Copyright Office guidance, can you fully 'own' something an AI generated entirely on its own from one prompt?

Yes, anything an AI makes for you is automatically yoursGenerally no — copyright protects human authorship; your own creative contribution is what can be protectedOnly if you paid for the toolOnly if you keep it secret

7. When is using AI on schoolwork actually cheating?

Always — any AI use is cheatingNever — using AI is always fineWhen you break the rules you were given, or present AI's work as your own thinkingOnly if the AI gets the answer wrong

8. A classmate wants to make a realistic AI fake of another student 'as a joke.' What's the right call?

Fine, since it's just a jokeNot okay — deepfakes of real people need consent and must never be used to harm, humiliate, or deceiveFine, as long as it looks realOnly okay if you don't post it publicly

9. An image tool keeps showing the same stereotyped person for 'a scientist.' Why — and what should you do?

It hates certain people; ignore itIt's repeating biased patterns from its training data; notice the stereotype and push against it on purposeIt's broken; stop using AIIt's correct and you should accept it

10. What is Content Credentials (from C2PA), and why does it matter for honest creators?

A way to hide that AI was usedAn open standard that attaches a tamper-evident record of where media came from and how it was edited — like a 'nutrition label' for contentA tool that makes AI art for freeA password manager