Challenges in adopting Artificial Intelligence and Machine Learning
Artificial Intelligence thrives on data. In recent times AI has been instrumental in creating new teaching and learning solutions that are now undergoing testing in different contexts. The beauty of any AI application is that it becomes more accurate when there is more accurate data available. Any AI application uses massive amount of data, and then the intelligence is built on it.
One very interesting aspect of the education sector is that the data that is available is accurate and also is built year on year. This happens both in the school ecosystem and also in the undergraduate and postgraduate education. Given that big data enables AI to reach its full potential, it would be fair to say that there is no data-driven AI without big data. This big data environment can be built at the country level if there is an infrastructure that is created by the government, which can give immense insights into the scenario of education in the country.
As on date, two areas have evolved in AI and ML with a special focus on education. They are learning analytics (LA) and educational data mining (EDM). Both of them overlap each other in terms of objectives and techniques. While EDM methods are drawn from diverse disciplines, including data mining, machine learning, psychometrics of statistics, information visualisation, and computational modeling, LA is more focused on learning content management systems and large-scale test results. LA combines institutional data, statistical analysis, and predictive modeling to identify which learners need help and how instructors can change academic behaviour.
While AI has great potential in the education space, India is a vast county, and the challenges that are thrown are unique and cannot be compared with other nations that have implemented AI and ML in education. I present the challenges that exist in applying and adopting AI and ML in the Indian Educational framework.
Lack of reliable high data: While the government has taken steps and initiative to collect data across higher education institutions, there is still a huge gap that exists between the ‘actual’ data and the data that is pulled by information systems.
Although the government launched National Academic Depository (NAD) with an objective to provide an online store that can hold all academic awards like the certificates, diplomas, degrees, mark-sheets, etc., duly digitised and lodged by academic institutions or boards or eligibility assessment bodies, it still has a long way to go and the work is still in progress. Unless there is a deadline that is set for all previous records to be updated, applying AI tools on this data will not be possible.
Lack of data at the district state and regional level: Many India educational still don’t have internal data that is available to the grassroots level. Many educational institutions still don’t have an ERP implemented, and hence, there is no central repository that collects data for further intelligence to be drawn from it. This is another challenge for implementing AI and ML, which can bring great insights and can help the teaching learning process.
The digital divide
India is a country that has great mobile penetration and also with new mobile players offering free data has made usage of internet reach tier -2 and tier -3 cities and towns across India. In-spite of mobile penetration at one end of the spectrum, there are schools that still lack basic facilities. There are engineering colleges that don’t even have digital projectors in the classroom and attendance being taken using the pen and paper method. One the other end of the spectrum there are schools which have unleashed the power of digital and automation solutions have been implemented to manage all the academic and administrative processes of the school or the higher educational institution to that extent that even Internet of things (IoT) applications are used for a specific process. This digital divide is a big challenge in implementing AI and ML.
While the applications of AI and learning analytics platforms can be used and the predictive algorithms can help teachers diagnose and anticipate learning difficulties faced by learners to implement personalised interventions to respond to those difficulties, are teachers equipped to use the same is a billion-dollar question? My personal experience has again made me see divergent scenarios. I have seen teachers who can’t even create and use a power point presentation in some schools and higher educational institutions, and I have also seen teachers who use learning analytics to find which student must be given focus based on the past and present academic records.
This is again a challenge to implement AI and ML across all educational institutions in India. Some policy changes have to created, and training must be made mandatory for teachers across all levels across the length and breadth of the country. Even if the government comes with the policy measure to train teachers, efforts have to be taken by teachers, and they must prepare, understand and grasp the new technological possibilities that digital and AI-powered tools can help education. MOOCs (Massive Open Online Courses) can be of great help in making this happen.