Luy IA? quiz

Sooy def quiz bi, tontu yi ak leeral yi dañuy feeñ.

1. Kiy waajal yoon dafay seet tukki yi mëna am ci ay sàrt yu leer. Lu lii di wane ci IA?
   • Bépp sistem IA dafa wara jàngee ci misaal yiñ etikete
   • Approche IA mën na jëfandikoo ay sàrt ak seetlu te du am reso neuronal
   • Interfaasu waajal yoon wi dafay firndeel xam-xam bu niro ak bu nit
   • Seetug sàrt yi dafay nuru tàggat ab xaajkat
2. Jàngum masin mooy ban diggante moo gëna am solo ak IA?
   • A subfield within AI that learns patterns from data
   • The hardware component that runs an AI system
   • A separate term for expert systems
   • An output format used only for text
3. Serwiisu yóbbug mbir yi dafay xayma waxtu yegsi gi ci done yi ci yoon wi ak ci dem bi ak dikk bi. Lan la jëfandikukat bi wara xam ci génne gi?
   • It proves the system understands every traffic situation
   • It is an exact schedule that cannot change
   • It is a prediction from inputs, not a guarantee
   • It only counts as AI if a neural network produced it
4. Benn ekip mingi tànn diggante ñaari jumtukaayi IA ngir liggéeyu jàppale bu leerul. Ban méngale moo topp tegtal bi?
   • Compare interface style and choose the more conversational tool
   • Evaluate only familiar examples because they are easiest to score
   • Choose the larger model before testing
   • Test fresh and ambiguous requests, and weigh the cost of errors
5. Organisatëru nataal dafay etikete bu baax nataal yiñ miin. Ban saytu bu ci topp mooy wax ci àppu biñ leeral ci tegtal bi?
   • Try unfamiliar or ambiguous pictures and review important errors
   • Assume every future picture will be labeled correctly
   • Treat familiar examples as proof that it understands all objects
   • Remove human review because the first results look convincing
6. According to the guide, why can fluent chatbot text still be wrong?
   • It checks each statement against sources before producing text
   • Its confident style shows the information was verified
   • It can produce likely-sounding text without checking whether claims are true
   • It only generates answers copied from one fixed encyclopedia
7. A writing assistant creates new paragraphs from a prompt. Which term describes AI that produces new content?
   • A system that only retrieves prewritten paragraphs
   • A rule-based template that fills fixed blanks
   • AI bu sosal
   • A classifier that chooses among existing labels
8. Before publishing an assistant’s paragraph with names and dates, what responsibility remains with the author?
   • Trust the paragraph if it reads fluently
   • Check that the names, dates and cited sources are accurate
   • Ask the model to add citations without checking them
   • Review only the paragraph’s tone