Based on Item Response Theory (IRT), CATs dynamically adjust their difficulty level according to the student’s responses to an earlier question and are generally associated with a relatively well-defined knowledge domain…writes Aarul Malaviya
Assessment is an integral part of a student’s learning journey. Even as it involves testing a learner’s ability to grasp classroom instructions and make timely progress, it is also in some ways a measurement of the teaching proficiency as an instructor, and the entry of AI has further changed the dynamic between the teacher and the taught.
Although the entry of AI in education was a force of disruption initially, it is in time turning out to be a remarkably transformative tool from the standpoint of both students and teachers. Indeed, AI can help teachers not only deliver a more customised and personalised learning experience for their students but also help a great deal in terms of students’ assessments.
While AI has been proving to be a great enabler for formative assessment of students, i.e., assessment and feedback during the ongoing delivery of instructions, it can also aid in making diagnostic and summative assessments of the student’s performance.
So, what are the different ways in which AI can help teachers with students’ assessments?
Can generate an array of tests
Teachers can employ AI very effectively to generate a wide array of tests and assessments for their students. These tests could account for the differential needs of individual students in terms of difficulty levels, the nature of the subject itself, adaptability, time and place, etc. For instance, a combination of AI and voice recognition can prove to be a highly effective tool for assessing the reading fluency of students in the context of formative testing. Since there is immediate feedback and the assessment and scoring can be done multiple times, teachers can spot specific areas where a student needs help and can make improvements thereafter. Importantly, based on large amounts of student data, AI can help in generating individualised assessments and tests.
Allows automatic grading of students
It is the sheer magic of AI drawing on its very nature that it can facilitate automatic grading and scoring of students on a given assessment in a given subject area. For example, AI-based automated essay scoring or AES uses natural language processing (NLP) to grade written essays and is used mainly in summative assessment. Although NLP goes back decades, the AI engines that underpin AESs have made tremendous progress today in terms of the sophistication of algorithms, data processing capacity, and overall efficiency. Similarly, Computerised Adaptive Testing or CAT has been in vogue for decades and involves automatic assessment and grading of students, again particularly used for summative assessment. Based on Item Response Theory (IRT), CATs dynamically adjust their difficulty level according to the student’s responses to an earlier question and are generally associated with a relatively well-defined knowledge domain.
Stealth assessments address exam-related anxiety
AI can also help teachers in creating and implementing stealth assessments and tests. Typically used in game-based learning, they can assist teachers in testing students without an emphasis on the process of traditional testing. As such, having been automatically embedded into the learning process and the content itself, these tests resolve issues related to students’ exam-related anxiety, one major factor not only impinging on the latter’s performance but also overall evaluation.
Helps early identification of at-risk students
With predictive analytics being at the core of AI applications, it allows an examination of the past performance of students based on their interest, aptitude, attention span, and task completion rates. Apart from the adaptive assessment that involves adjusting the difficulty level for individual students, AI also supports teachers in the early identification of at-risk students as well as identifying problems that would likely require a remedy.
Aids surveillance & prevents cheating
Even as AI is increasingly being cited as a tool that ‘facilitates’ students’ cheating, especially in light of the ever-rising popularity and usage of Chat GPT, the generative large language model tool,it can also be used as an instrument of surveillance and indeed for prevention of cheating. Today, AI-powered remote proctoring tools are being developed that also include audio proctoring, video proctoring, and image proctoring, especially making use of facial recognition technologies. In addition, various plagiarism software focused entirely on AI-generated content is making their presence felt among users including teachers.
Frees up teachers’ time
The increased availability of AI-based assessments and tests also allows the teacher to shift back and forth between testing and teaching with more efficiency and more frequently thereby improving overall learning outcomes in the end. The bottom line is that AI applications greatly reduce the workload of teachers allowing them to concentrate better on their pedagogical method, instructional content, and the quality of delivery itself.
Therefore, AI can indeed make the assessment process more accurate, personalized (for the student), and swift for the 21st-century teaching community. In a time and age of increasingly digitized curricula, the advent of advanced machine learning-based assessment systems that continue to make themselves better while being used would further make it easier for teachers to assess and evaluate their students.