CELPIP

CELPIP Listening Part 6: Viewpoints

Practice Test 6

Part 6. Instructions.
You will hear a talk given by ONE person where different viewpoints on a single issue are presented. In the real test, you hear the discussion only once.

After the discussion, you will hear six questions. Choose the best answer, (A), (B), (C), or (D).

Try to answer after one listen to simulate test conditions.


Automated Essay Marking in Canadian Universities

You will hear a commentary about the use of artificial intelligence to mark student essays at Canadian universities. Listen carefully and answer the questions.

1. The talk is mainly about ___.

    how artificial intelligence is changing university admissions

    different views on using machines to mark student essays

    the history of essay assessment in higher education

    the cost of running Canadian university programs

2. According to Dr. Tremblay, automated marking systems can ___.

    assess the structure of an argument, not just its surface

    operate without any human oversight at all

    replace classroom instruction in writing courses

    be customized for individual student needs

3. Professor Donnelly's central concern is that ___.

    the systems contain biases against certain writing styles

    students should be allowed to choose their marker

    marking essays is itself part of teaching the student

    machine marking is slower than its supporters claim

4. The Canadian Federation of Students is calling for ___.

    more funding for human markers in large courses

    student representation on university technology committees

    clear information about how and when the systems are used

    an immediate ban on automated marking systems

5. Unlike Donnelly, Tremblay would most likely argue that ___.

    machine marking is mostly suitable for short answers

    the systems work best when used as a backup to humans

    human markers should be removed from the process entirely

    feedback speed matters as much as feedback quality

6. The speaker's main conclusion is that ___.

    the Federation's demands should be rejected as unrealistic

    the debate ultimately turns on what universities are for

    automated marking should be adopted with caution

    Tremblay's research has settled the question


View Transcript

Good afternoon. The subject of today's talk has moved very quickly from the margins of academic conversation to the centre of it: the use of artificial intelligence systems to mark student essays. Several Canadian universities have begun piloting these tools, and a handful are now using them at scale in first-year courses. The debate has tended to generate more heat than light. What I want to do this afternoon is lay out three distinct positions and then offer a tentative judgement of my own.

Let me start with Dr. Felix Tremblay, an education technology researcher at McGill. Tremblay has been one of the most prominent academic voices in favour of automated marking, and his case rests on three claims. First, he argues that the consistency of machine marking is genuinely superior to human marking, which varies depending on time of day, fatigue, and the marker's mood. Second, Tremblay points out that students receive feedback within hours rather than weeks, which his research suggests has measurable benefits for learning. Third, and most controversially, he maintains that the systems are now sophisticated enough to recognize argument structure, not merely surface features like grammar and length. His position is that resistance to the technology is rooted more in professional anxiety than in a clear-eyed assessment of what it can do.

A sharply different view comes from Professor Maeve Donnelly, who chairs the English Department at the University of Victoria. Donnelly is not a technological pessimist, and has been careful to distinguish her position from a blanket rejection of digital tools. Her concern is more specific: she argues that the act of marking an essay is itself a form of teaching, and that delegating it to a machine fundamentally changes the relationship between an instructor and a student's writing. According to Donnelly, what is lost is not measurable in the data the systems collect. A skilled human marker notices the student behind the prose, the moment when a writer is reaching for an idea they cannot quite articulate, the early signs of intellectual independence. None of that, she insists, can be captured by a model trained on patterns in past essays.

The third voice belongs to the Canadian Federation of Students, which has staked out perhaps the most pragmatic position of the three. The Federation is not opposed to automated marking in principle. Its members acknowledge that human marking has its own well-documented problems and that students have long complained about inconsistency. However, the Federation argues that the current rollout has been driven by administrators looking for cost savings, with insufficient consultation of students and instructors. Their demand is for transparency: students should know when their work is being marked by a machine, should have a right to appeal to a human marker, and should be told how the system was trained.

My own view is that Donnelly captures something important that Tremblay's enthusiasm tends to obscure. Marking is teaching, and the relationship between a writer and a careful reader is not merely an inefficient process to be optimized away. At the same time, Tremblay's data on feedback speed and consistency cannot be dismissed. The Federation's call for transparency strikes me as the right starting point, but it will not resolve the deeper question of what universities are for. If they exist to certify competence at the lowest cost, automated marking is reasonable. If they exist to develop minds, the case is more complicated than its proponents have acknowledged.

© 2001-2026 esl-lounge.com