11. The Too-Perfect Email
Which phrase is the biggest AI clue in a casual email?
Recommended answer: B. It sounds polished but unnatural for a quick note.
More fast AI tests for weird details, overconfident answers, and prompts that can go sideways.
Which phrase is the biggest AI clue in a casual email?
Recommended answer: B. It sounds polished but unnatural for a quick note.
An answer says 17 plus 28 equals 46. What should you notice?
Correct answer: A. The correct sum is 45, a simple check that models can still miss.
The output shows a laptop floating in a forest with glowing code in the trees.
Correct answer: A. The forest, laptop, and code all point to a stylized tech prompt.
Choose the weaker answer to "What should I cook tonight?"
Recommended answer: B. It avoids the concrete suggestion the user asked for.
A chatbot gives a source that sounds real but has no author, date, or link. What is the issue?
Correct answer: A. Missing source details can be a sign the reference was invented.
Which prompt is more likely to produce visual glitches?
Correct answer: B. Many repeated bodies, reflections, and geometry increase the chance of errors.
Which sentence sounds like a model trying too hard?
Recommended answer: B. Buzzword stacking often feels less human and less useful.
A prompt asks for "a dog beside a car" but the result crops off the dog. What would help?
Correct answer: A. Composition instructions reduce accidental cropping.
Pick the more natural product review.
Recommended answer: A. Specific tradeoffs feel more credible than vague praise.
If the prompt says "make the logo pop," what could AI misunderstand?
Correct answer: A. Figurative language can be interpreted too literally.
Some challenges are about tone or design judgment, so the answer explains the strongest clue rather than a strict rule.
No. Every page is a separate static HTML file linked with normal anchor tags.