Cuestionario vinculado a una guía · duro Nivel

Prueba de calificación y comentarios de IA para profesores

Prueba indicaciones de retroalimentación basadas en rúbricas, detectando evidencia inventada y desviación de puntaje, y preocupaciones de equidad, incluido el sesgo del detector de escritura de IA.

Rutas de guía relacionadasCalificación y comentarios de IA para profesores
Pregunta 1 de 8

According to the guide, which AI grading job is safer and more useful?

Vista previa de cada pregunta8 preguntas

Las respuestas y explicaciones aparecen a medida que realiza el cuestionario.

  1. According to the guide, which AI grading job is safer and more useful?
    • Assigning final report card grades
    • Drafting rubric-based feedback that the teacher reviews
    • Deciding academic misconduct cases
    • Ranking students against each other
  2. Which accuracy problem involves the AI praising sentences not in the student's work?
    • Score drift
    • Length bias
    • Invented evidence
    • Rubric overfitting
  3. What does the guide say about AI writing detectors?
    • They are unreliable, and some disproportionately flag non-native English writers
    • They are fully accurate for essays
    • They are required by FERPA
    • They only flag plagiarism from books
  4. How does Gradescope's answer-grouping feature work, as described in the guide?
    • It assigns final grades automatically with no teacher input
    • It writes essays for students
    • It detects cheating using webcams
    • It groups similar answers so the teacher grades each group once
  5. What automated check catches invented evidence in AI feedback?
    • Counting the words in the feedback
    • Confirming that every quoted passage actually appears in the submission
    • Running a spell-checker on the rubric
    • Checking the file size
  6. What is calibration in the context of adopting an AI grading tool?
    • Adjusting the screen brightness
    • Letting students grade the AI
    • Scoring papers teachers already graded and comparing results criterion by criterion
    • Increasing the randomness setting
  7. Why does the guide say AI grading is not objective just because it is automated?
    • It applies patterns learned from data, including that data's biases
    • Computers cannot count
    • It only works in English class
    • It is always stricter than teachers
  8. For which type of work does the guide say AI feedback works best?
    • Final exams
    • College admission decisions
    • Suspension hearings
    • Frequent, low-stakes formative feedback