What is the main risk when AI makes question generation faster?, Weak item-quality problems can be multiplied at scale., Teachers can no longer write questions., All distractors become too easy., Assessment becomes automatically valid., Why should generation and judgement be separated?, Judgement is unnecessary after generation., AI can draft, but a human must audit construct alignment, evidence and quality., AI cannot produce text., Humans should never use automation., What does source-bounded prompting require?, Removing the original text., Allowing invented details to fill gaps., Supplying the source/rubric and constraining output to that evidence., Asking for any plausible content., What should AI do when source information is insufficient?, Create a reference., Guess silently., Change the construct., Mark uncertainty rather than invent a plausible answer., Which assessment use requires the strongest human moderation?, A score or feedback decision that affects learner progression., A private brainstorming list of possible stems., A draft worksheet title., A teacher’s personal note., Which action demonstrates AI evaluation competence?, Use the AI key without checking., Challenge a distractor and revise it using source evidence., Press 'generate.', Copy all items to a test., What should be verified in an AI-generated multiple-choice item?, Only the number of options., Only response time., Stem, key, distractors, level and source evidence., Only formatting., Which data practice is safest?, Upload full names and scripts to any public chatbot., Store test banks publicly., Share learner scores in open links., Use anonymised or non-identifiable learner data in approved systems., What is a test-security concern?, Exposing secure item banks or answers to unapproved services., Using a rubric., Providing feedback., Checking distractors., How should AI-generated rubric wording be used?, As a scoring decision by itself., As a draft checked against the construct and actual performance criteria., As final criteria without review., As a replacement for the learning objective., Which feedback comment is strongest?, A score with no explanation., A claim about learner ability based on one error., One that is criterion-linked, evidence-based and actionable., A fluent generic paragraph unrelated to the learner text., What should the teacher do with an implausible distractor?, Keep it because AI generated it., Make it the correct answer., Hide the source., Revise or replace it because it will not diagnose misunderstanding well., Which prompt is more defensible?, 'Using only this passage, write items and cite the evidence for each key.', Write 20 hard questions about anything related to the topic., Invent details if needed.

QUICK QUIZ

tarafından
Yazdır
Yerleştir

Skor Tablosu

Görsel stil

Seçenekler

Yapay Zeka Geliştirilmiş: Bu etkinlik, yapay zeka tarafından üretilen içerikleri içerir. Daha fazla bilgi edinin.

Şablonu değiştir

Otomatik olarak kaydedilen geri yüklensin mi: ?