Publish at April 17 2024 Updated April 17 2024
Of all a teacher's tasks, grading is probably one of the most complex. The sheer scale of the task can be daunting, especially at secondary and higher education levels, where the number of papers to be processed can rise from thirty to hundreds. The complexity increases with the type of work. It's easier to correct a multiple-choice questionnaire than an open-ended question, or one that requires you to analyze your approach.
Could artificial intelligence (AI) support teachers in this task? As technology gradually makes its way into all sectors of education, many believe that generative AI (such as ChatGPT) could transform assessment. The question is, how?
The algorithm knows nothing about fatigue, and is generally not biased towards learners. As a result, it can quickly correct a copy, note the score and pass it on. A process that takes teachers hours could then be carried out by the machine so that they can concentrate on other tasks.
What's more, while for the moment the artificial intelligences in place are mainly specialized in word processing, others are working on voice recognition. In this way, an AI could analyze a student's reading fluency, for example, or potentially assess the pronunciation of a foreign language.
This process would also enable students to get feedback much faster, and know whether they're on the right track or not. The teacher would be able to see right away the cognitive errors made and help them to correct the data or the process needed to solve a problem. The algorithm would thus become an ally of the teacher in working on learning difficulties.
The whole question of personalization lies at the heart of the implementation of artificial intelligence in assessment. Since it is capable of immediately detecting a person's faults and strengths, it would be able to adapt test content according to this data. It would spend little time on what has been mastered and focus on what is causing problems, providing exercises, resources and additional explanations to help the learner's understanding.
This also comes with current questions and limitations. Obviously, data acquisition raises a big issue: how do we ensure that data remains private and secure? Particularly as we know the extent to which companies are snapping up information for commercial use. Artificial intelligences used in education should therefore guarantee that information is impossible to obtain.
Another problem is the quality of the information possessed by the algorithm. A less robust and poorly tested model can lead to erroneous assessments. There is also the possibility of bias depending on the database used. This has been seen in the use of AI in the courts in the USA, where it reproduced certain racist decisions. We therefore need to ensure that this type of judgment is not present, and above all to put in place a structure so that humans can step in and rectify these potential flaws in order to offer a fair assessment.
There's no doubt about it, the advent of artificial intelligence is transforming school assessments. But while algorithms can easily create and correct questionnaires, they are not necessarily best placed to assess skills such as creativity, resourcefulness and others.
Thus, the use of portfolio assessments or scenarios requiring learners to think critically or be able to act in simulated events remain in the far more human realm. This will give the school system the opportunity to better prepare learners for situations they will encounter in the future.
As more and more schools consider the use of AI in assessment, it's important to remember that it's not a magic solution. It needs to be thought through so that there is room for the human element. Algorithms can make mistakes, and it's up to teachers to revise if this has been the case. It's also an opportunity for teachers to implement other, more creative and meaningful evaluation approaches.
Image: limbi007 / DepositPhotos
References:
"5 ways artificial intelligence might transform assessment practices." John Spencer. Last updated: October 2, 2023. https://spencerauthor.com/ai-assessment/.
"AI in higher education: impact of AI on student assessment." Tensorway - AI Development Company. Last updated May 4, 2023. https://www.tensorway.com/post/ai-for-student-assessment-in-higher-education.
Britten, Jody. "AI and assessment in education: learning from Australia." Medium. Last updated December 6, 2023. https://medium.com/@team4tech/ai-and-assessment-in-education-learning-from-australia-67fc8f8c2999.
Bryant-Aird, George. "Generative AI can change assessment for the better." Wonkhe. Last updated April 11, 2023. https://wonkhe.com/blogs/generative-ai-can-change-assessment-for-the-better/.
Hamilton, Laura S., and Kamal Middlebrook. "How AI can help create assessments that enhance opportunities for all students." The 74. Last updated January 23, 2024. https://www.the74million.org/article/how-ai-can-help-create-assessments-that-enhance-opportunities-for-all-students/.
"How AI is changing assessment in higher education." LinkedIn. Last updated August 10, 2023. https://www.linkedin.com/pulse/how-ai-changing-assessment-higher-education-wixhighered/.
Matthews, Ed. "How can AI tools improve student assessment outcomes?" CAT. Last updated October 30, 2023. https://www.taotesting.com/fr/blog/how-can-ai-tools-improve-student-assessment-outcomes/.
Neendoor, Snehnath. "AI-enabled assessment: Redefining evaluation in education." hurix digital. Last updated April 2, 2024. https://www.hurix.com/ai-enabled-assessment-redefining-evaluation-in-education/.
"Rethinking assessment strategies in the age of artificial intelligence (AI)." Charles Sturt University. Last updated February 28, 2024. https://www.csu.edu.au/division/learning-teaching/assessments/design-standards/rethinking-assessments.
Zalli, Detina, and Aurela Ramaj. "The future of student assessment in the age of AI and ChatGPT." Higher Education Digest. Last updated October 19, 2023. https://www.highereducationdigest.com/the-future-of-student-assessment-in-the-age-of-ai-and-chatgpt/.
You might also like:
The mystery of animal onomatopoeia
Have you ever wondered why animals sound different in different languages? Could it be that the French duck goes quack quack while its English counterpart goes kwak-kwak? Perception or misunderstanding? Do you give your tongue to the cat? Come and unravel the mystery of animal onomatopoeia!
Are values more important than tasks?
The question of values has always had a certain importance in the workplace, but this has been confirmed in recent years. A series of factors have meant that Generation Z, among others, no longer accepts just anything when it comes to employment.
Reasons for dropping out and hanging up
Learners who drop out are often perceived as failures by the school system. While the situation has not improved with covid, it should not be forgotten that a significant percentage of dropouts eventually return to school. By interviewing them, it is possible to understand what led to dropping out and hanging up.
A "robustness pedagogy" is currently being invented. It observes the lessons of the living world, which knows how to dispense with centralization and outrageous performance in favor of the circulation of energies and emotions, and adaptation to what comes next.
Since Bachelard, Serres and Astolfi, we know that teaching is not filling an empty box, nor replacing one knowledge with another. Old representations resist! Teaching is often a matter of bringing out what needs to be overcome, understanding what sticks in order to finally teach the knowledge and know-how useful to learners. Actions based on the concept of "epistemological obstacle" provide us with useful leads.
