The impact of AI on education is not a new concept. From online education to the increase in building edtech, AI has been transforming education at various levels. Some AI-powered chatbots act like personal tutors for students, capable of answering common questions and doubts at their disposal. Using NLP, these bots are designed to address child's queries as and when they arise, learning to better understand and clarify concepts.  

AI also offers customized lessons based on the aptitude of every child. It also provides regular and efficient assessments and real-time evaluations. In addition, language translation powered by AI makes every model inclusive. 

Already a widely-adopted solution, AI in edtech has crossed the inflexion point owing to the lockdown-induced acceleration of digital learning. Virtual modules of access to knowledge and understanding are now servicing an even larger volume of students. The government has also stepped up to ensure the tech-led pedagogy's preparedness for the pandemic's imposition.

Now, AI has taken a new toll. It can now predict the grade that a student might obtain.  

Grade prediction 

The initial work focusing on grade prediction was based on applying a series of pre-established rules to a relatively simple set of facts. More recently, however, the results proposed to analyze the entire record of a student's interaction with her educational platform and use a complex neural network to achieve that grade prediction.  

For instance, in the work presented by Alonso-Misol et al., the performance of different algorithms is compared, obtaining an accuracy of 96% when predicting the grade of an exam. Moreover, in 96 out of 100 students, grades assigned by the teachers were close to the algorithm's predictions. 

A distant dream 

Exams are the most hated activity by students. Though the prediction is successful, it will only produce meaningless numbers if the test disappears. 

Predictive systems rely on supervised learning techniques. The data from the current course is analyzed and compared with data from past courses. If there is a pattern of activity in past courses related to the achievement of a certain grade, then students who have that pattern in the current course will be predicted to have that grade. In other words, a student will likely obtain a similar grade to that obtained by students who had a similar interaction with the platform. 

Predictive systems will be successful to the extent that the course being analyzed performs equivalently to previous editions of the same course.  

If there is a month-long course in which students have to submit an activity on Friday of each week, there will be students who only generate activity on Friday to make the delivery. Students will also generate activity throughout the week, with more intensity on Friday. 

If the teacher decides that the activities corresponding to the four weeks are all due at the end of the month, the pattern of students would again change, and there will be students who might not enter the course until last week.  

The change in teaching methods will affect the predictability of the models. The eradication of examinations in schools might even lead to instances such as the UK’s A-level exams, which was a classic case of AI-bias. Similarly in 2020, the use of algorithms to replace exams hit the headlines when it was shown that they penalized students from state schools and low-income post codes. 

Therefore, cancelling exams will still be a distant dream for our students. However, the overall objective of this prediction system is to detect students at the risk of dropping out to offer them adequate support. They are also helpful in anticipating the resources that will be needed.