India, being the fastest growing economy with the second largest population in the world, has a significant stake in the AI revolution. AI refers to the ability of machines to perform cognitive tasks like thinking, perceiving, learning, problem solving and decision making. AI is a constellation of technologies that enable machines to act with higher levels of intelligence and emulate the human capabilities of sense, comprehend and act. The truly transformative nature of the technology, yet the nascent stage of its adoption worldwide, provides India with an opportunity to define its own brand of AI leadership.

Challenges in improving the quality of education:
Factors that pose challenges to improving the quality of education:
a) Multi-grade and multi-level classrooms: For a large proportion of schools, especially in small or remote villages, it is not viable to have separate classrooms and teachers for different grades / classes. Consequently, the teacher is faced with a heterogenous group of children in the same classroom, with wide variations in their classes, ages, abilities and learning levels. This large variation poses a huge challenge to the teacher and is a common cause of poor teaching-learning, thus leading to poor learning outcomes.
c) Lack of interactive pedagogy and ineffective remedial instruction: Teaching-learning processes in most classrooms are highly rote-based and non-interactive. Remedial instruction, where conducted, typically lacks customisation to the child’s learning level, abilities, and pace of learning.
d) Inadequate attention / action for likely drop-outs: Several children may be at risk of dropout due to various factors, such as inadequate school infrastructure, poor teachers, poor school readiness, language barriers, large learning gaps with respect to grade level, family circumstances (e.g. migrant families), poor nutritional or health status, etc.
e) Large teacher vacancies due to uneven distribution across locations: Large number of teacher vacancies are mostly not due to an overall shortage of teachers in a State – instead, they are due to uneven distribution across different geographical areas within the State. For instance, recent figures for Uttar Pradesh revealed 1.74 lakh teacher vacancies at elementary school level, but a simultaneous surplus of 0.66 lakh teachers across the state.
f) Professional development courses / training do not cater to real needs and have poor coverage Existing teacher training is typically a generic kind of an exercise. It is not linked to the specific weaknesses / requirements of a teacher – for instance, a teacher with poor arithmetic understanding requires corresponding training to clarify arithmetic concepts. Consequently, most teacher training exercises end up as wasted public expenditure, with little or no benefit to the teacher and her / his students. Similar issues exist with respect to training of other staff such as school headmasters/principals. The coverage of existing training programs is also extremely low, typically less than 20% annually.
g) Low adoption of existing technologies: A recent survey found that level of adoption of technology in schools is lacking, and can be largely attributed to lack of teacher training, despite provision of the ICT infrastructure. While 83% of the teachers surveyed use computers, the use is limited primarily to audio / visual display, or student practice. A meagre 41% and 27% use technology for tracking student data and participating in forums respectively.

Role of AI in education:
Several AI tools are being successfully used in other parts of the world, and they can be adapted to the Indian context to target specific challenges.
a) Adaptive learning tools for customised learning: While AI may not completely replace a teacher, it has the potential to greatly assist teachers in efficiently and effectively managing multi-level / multigrade classrooms, by judging learning levels of individual students, and allowing automated development of customised educational content adapted to each child’s class and learning level. Assessing time spent by a student on each part / page of the learning material, for example, would allow real-time feedback on student performance to help the teacher appropriately tailor her guidance to the child. This concept can be extended to automatic grading of tests, as well.
b) Intelligent and interactive tutoring systems: Intelligent Tutoring Systems can provide great benefit to students through delivery of learning materials adapted to the child’s proficiency level, learning style, and pace of learning. In-built pop-up questions tailored to students, for example, can help increase interactivity, and catch student’s attention and interest. It can also help in assessment of student’s level of attention or comprehension to appropriately design remedial instruction. GradeGuardian, for example, uses predictive models and visualisations for student performance with an interactive dashboard showing anticipated effect of policy changes.
c) Predictive tools to inform pre-emptive action for students predicted to drop out of school: Analysis of test results and attendance records using AI can be used to predict probable student activities and inform pre-emptive action. For instance, in a recent preliminary experiment conducted in Andhra Pradesh, AI applications processed data on all students based on parameters such as gender, socioeconomic factors, academic performance, school infrastructure, teacher skills, etc., with the objective of helping the government identify students likely to drop out. Test results could inform suggestions to enroll students in vocational studies. Additionally, redressal mechanisms could be put in place to identify students whose performance can be improved by focus of existing schemes to their family.
d) Automated rationalisation of teachers: AI tools can be used to develop automated teacher posting and transfer systems, using analytics based on demand – supply gaps across schools in the State, candidate’s prior postings, candidate preferences, etc. This would help in plugging of gaps in teacher distribution more effectively.
e) Customised professional development courses: To tackle issues of poorly designed professional development courses with poor coverage, adaptive AI tools can be used to design automated, customised professional development training content for the teacher based on their performance, identification of their knowledge and skill gaps. This could then be continuously adapted as teacher’s skills and concepts improve.