Introduction
From information age, we have now moved towards the age of data gathering and data interpretation. From the plethora of information and data, we have to extract some tangible ‘meaning’ for interpreting a trend or forecasting a future; these have become the realm of artificial intelligence or AI. As a scientist, I have always used data to interpret a hypothesis or an assumption before the start of an experiment. We would call it statistical interpretations then, but times have changed and AI is in fashionable phrase today!
AI 101
AI is here to stay and is rapidly impacting everything that we do! AI is defined as machines or devices that think and react like humans – machines should have cognitive abilities and have attributes such as problem-solving, perception and planning. The feasibility of incorporating these abilities and attributes in machines can be quite difficult, and requires immense about of programming efforts with analysis.
The easy examples of AI are Siri and Alexa and their voice-enabled services, or self-driving cars, planes on auto-pilot, etc. IBM’s Deep Blue chess player and the Debater are fascinating examples of AI. Interestingly Tesla cars use AI to avoid accidents (one of my friends was saved by Tesla’s AI. When a car in front hit a divider and summersaulted……….her car’s sensors and AI used an algorithm to detect the possibility that the crash was about to occur, and her car was brought to a halt automatically!!) This is a fantastic new application of AI isn’t it?
Thus AI when used wisely can have a range of useful applications indeed!
AI in Education
So how does AI compliment and supplement our teaching and learning endeavours? AI must be applied intelligently for a ‘smart’ impact on the learning outcomes.
We know that the explosion of resources on the internet has given vast learning opportunities to millions of eager learners who previously did not have access to good books or even teachers – I am especially referring to paucity of educational resources in India. The internet has opened up this previously narrow view of education and has opened up a smorgasbord of offerings to everyone who has access to the internet (although one has to be a bit discerning about sites the students are learning from, as some of these can be quite misleading and incorrect).
Admin automation (ERP) along with LMSes and MOOCs are already being used widely (unfortunately not much in India, we are so much behind this ed-tech curve) in many educational institutions. Automated attendance, grading, scheduling, analytics, school management issues can be performed routinely with AI.
Learning management systems (LMS), massive open online courses (MOOC) use data to provide ‘just-in-time’ like learning assistance to students (and teachers as well). These systems can be used to monitor the teaching-learning processes very effectively.
Adaptive learning
Automated tests, feedback and grading are common form of application of AI in education. Let me give you an example here: a student took a test and the result is as shown below:
The course LMS/exam engine will direct the student to learn and practice more in modules with low marks! The student can go back and forth many times till the skill levels are acquired, and then appear for exams, or may be proceed for an internship or hands-on projects/experiments, etc.
But for machine AI to achieve this – the person writing the code has to be absolutely thorough in gauging what is required from the learners, that is know the learning trajectory correctly to bring about the desired results from students, keep them engrossed and motivated in the learning process; this is not an easy task! Vast amounts of information have to be provided to the machine to offer learning paths to students. If properly harnessed, the AI’s potential can be immensely beneficial to the learners as well as the teachers (not to forget the stakeholders and leaders of an educational institution).
I was a community TA for a HarvardX course on the edX platform, where adaptive learning was used in the version 2 of the course (please see: https://vpal.harvard.edu/files/vpl/files/paper_167.pdf). It was clear that adaptive learning enabled the learners to score higher in their tests. The HarvardX group along with TutorGen used an adaptive engine with a strategy in the form of an algorithm based on Bayesian Knowledge Tracing (probability calculations).
Another aspect that may be in vogue in due course of time is tutoring support – AI can be used for tutoring support outside the classroom too. This would indeed get rid of sub-standard tutoring and coaching classes especially in India, and give the students what their class teacher has recommended for them to read, learn and practice. (This could prove disruptive, which is a good thing!)
Summary
AI in education does not mean robots stomping around and replacing teachers!!
AI is here to stay, and AI in education is a must for the various teaching-learning paths and processes.
The leadership in education needs to chart a course for assimilating AI in their institutions so that it can positively impact ‘smart’ learning outcomes of a module, of a course, and of a complete program as well. Just like IBM’s Watson is using AI in healthcare, offering diagnostic assistance to doctors, similarly AI in education provide a diagnostic tool for all stakeholders as to what is actually happening in the classrooms – is the teacher teaching, is the learner learning? And when this gap is bridged, we would indeed have engineered a revolution in education!
Read more
You may read my other blogs relevant to this post:
https://atabhagwat.wordpress.com/2014/10/27/diagnostic-formative-and-summative-assessments/
https://atabhagwat.wordpress.com/2014/07/08/quality-assurance-in-education-and-training/