AI Solutions That The Govt Education System Could Benefit From

Raj Reddy Centre for Technology and Society (RCTS) recently organised a closed roundtable discussion on the enablers, barriers and opportunities that exist for the adoption of technology in the Telangana Government education system. Here’s a summary of the deliberations.

While the pandemic put the spotlight on digital educational content, there are many other advantages to technology in education. Apart from online course content, the Telangana Government has over the last few years been investing in the development of technology in education through the creation of various software and custom applications for functional tasks such as teaching tools, digitization and administration. The aim of all these investments has been to streamline education in the state and provide students with the most optimal learning resources possible. The efforts albeit commendable, are not without hurdles. Hence a concerted brainstorming endeavour by relevant stakeholders and key collaborators such as teachers, headmasters, school complex heads, Mandal Education Officers (MEOs), District Education Officers (DEOs), Government department liaisons and grassroot NGOs was facilitated by the Raj Reddy Centre for Technology and Society (RCTS) at IIITH. The RCTS is an initiative of IIITH to enable research and emerging technology-led solutions for grassroot education and public health, with a specific emphasis on rural areas.

Tech Resources
In order to set the context of the current government education system in the state, the congregation made a list of the technologies and software in use, their advantages as well as the drawbacks and loopholes. For various administrative tasks such as attendance entry, mid-day meal tracking, grade tracking and so on, multiple software modules have been generated that collectively function as an Integrated School Information System. Applications such as Excel, Google Sheets, Google Drive, Dropbox, etc. are the primary platforms where teachers and administrators perform their day-to-day work and collaborations. This has greatly sped up their workflows in many cases. While a large extent of the work communication relies on Gmail, in recent years, a visible transition to WhatsApp has occurred. Technology resources are being made available to students through the disbursement of gadgets like smartphones, smart tablets, and even computers through other support programs such as NGO interventions and more. The backbone for all of this which is a stable internet connectivity is also being ensured through the laying of T-Fiber connections in most government schools at the district level.

Drawbacks Of Current Systems
There are infrastructural issues that are related either to hardware and software resources such as dilapidated labs and equally old computer systems, non-standardised procurement of hardware due to support programs of various NGOs, software modules developed for specific purposes having long outlived their use, software not undergoing regular updates leading to system crashes, and data loss. In addition to these, despite the laying out of T-fiber lines to many schools, the equipment needed for hooking on to the internet is missing or in an unusable state. This leads to a high dependency on mobile data for connectivity both for students as well as teachers and administrators which in turn results in poor adoption of any technology initiative in rural areas, where signal strength and coverage issues persist. Another infrastructural issue relates to that of manpower – lack of dedicated personnel, either as faculty or Systems staff to train students and to handle the IT needs respectively. In fact the lack of dedicated employees to handle administrative tasks is proving to be a distraction for teachers.

Issues related to technology can be classified in two ways; those related to the design, development and deployment of tech solutions and those concerned with lifecycle planning. In the former, software modules designed on an ad-hoc basis with little or no attention to compatibility or existing applications lead to redundancy, lack of consistent developer teams contributing to the said redundancy,  poor data practices leading to cumbersome workflows, lack of assistive features in the systems compounding the drudgery of every-day tasks and so on. In the case of lifecycle planning, many issues were highlighted such as the lack of attention to bugs making the software unsafe and unusable, the collected data either not stored properly or ineffectively processed and utilized. Reporting of the data was found to be a one-way street with only teachers and headmasters sharing data of either attendance, grades etc to their District Education Officers and no analytical insights being fed back to the teachers. Plus, since long-term milestones have not been established for assessing the usefulness, and effectiveness of the various technology solutions, there is no scope for improving upon them based on planned road maps, user feedback, and current situations.

Lack of awareness of technology tools and their perceived value due to gaps in clear communication on all the available tools, and the absence of an understanding of the long-term impact of digitization, is greatly affecting the adoption of new technologies by on-ground personnel. New systems that are being built have minimal to no end-user input taken during development, alienating the end users’ interests and concerns. Training of end-users is inefficient coupled with a dearth of support documentation and user guides to assist them in adopting the digital tools in their daily workflows.

Where Data Can Be Digitized
With the large governmental push towards the digitization of data in the field of education, student grades, student attendance, and teacher attendance is being digitized in some fashion or another. Information such as “Unified District Information System for Education” (UDISE) codes for schools, information regarding mid-day meal programs, and other such support programs are also being digitized. Besides these, the panel was of the view that in order to track student performance throughout the journey in a government educational system, digitization of the following is critical: the creation of a centralized education registry and integrated school information system where with a single unique identifier code (possibly with existing identifiers like Aadhaar) is generated for each student, teacher and administrator to track their information much like the UDISE code for schools. This will help    reduce some of the repeated data entry operations required to work with multiple software solutions, increase the reliability of data storage by enabling centralized monitoring and control, enable a better understanding of why a given tool is not being used to the expected extent, in turn fostering a better functionality design for future systems.

While information on student registration and their Transfer Certificates is partially digitized currently, there are some gaps that persist. Along with maintaining online records of new students, old student information ought to be converted to the digital format too. Student grades and attendance records should be digital too enabling teachers to get the bigger picture of students’ performance over the years whilst comparing it with other students too. In order to facilitate online learning, existing textbooks, and other relevant study materials along with newer audio/video content customized to suit respective learners’ needs should be made available on online platforms. The infrastructure resource information in terms of equipment, labs and so on must be made available online in order to track it periodically and ensure optimum utilisation.

AI-Based Solutions That Can Help
Based on the complexity of the solution design and implementation, the panel proposed solutions that could be graded in terms of low, medium and high. Solutions that didn’t require any major hardware or infrastructural investment or manpower deployment were categorized as ‘low’ in complexity since they would build upon existing resources and function in the interim while long-term AI solutions would kick in. These include but are not limited to, standardized pivot tables for use with Google Sheets or Excel to provide quick analytics, downward communication of the said analytics to motivate teachers in data collection and aggregation, centralized systems for the Telangana Government Education Department such as billing to ease tracking of supplies in govt-supported programs across the state, automated data validity checks during data entry, digital access and authorization for educational officers to remote work portals to enable them to work from any location thereby increasing productivity.

Solutions that are of medium complexity involve standardization of reports and data formats across different platforms and government departments, standardized workflows for periodic data validity checks and student record updates, automated generation of personalized student performance reports, informative dashboards for quick summarization and visualization of information to the users, highlighting students, teachers, and schools in need of intervention based on various set criteria such as performance, attendance, infrastructural requirements, and so on, development of a learning recommendation engine to suggest learning paths based on student historical performance and automated performance rewards for teachers through preset criteria, obviating the need for manual intervention.

The creation of AI-driven tools to facilitate many of the administrative as well as learning-led tasks assumes a high level of complexity. It includes the creation of digitization tools to scan older records, the building of various Indian language corpora and machine translation engines to allow for easy processing of historic records and increased user accessibility, automated attendance systems for tracking participation in school and government support programs using student/teacher biometrics, automated assessment of student attention and performance using computer vision-based systems and the creation of all these tools that can support multiple languages. The latter would be immensely valuable in the context of our multi-lingual nation by making content available in the native tongue of the user and in the case of digitizing historic records which may be in different languages.

Points To Ponder
The panel was of the opinion that while AI-based solutions are a boon and can greatly assist especially in the creation of digital content and so on, priority must nevertheless be given to physical interaction for teaching. While some issues may require novel solutions, for simple issues solutions that fully utilize existing tech resources must be proposed. The outreach and training programs ought to be planned in conjunction with new solutions developed to better equip all the end-users. In fact the end-users must be kept in mind at all times and their appropriate feedback taken in the development process. Finally, solutions must be rolled out gradually with pilots in order to catch issues early and avoid failures arising from large-scale disruptions.

Sarita Chebbi is a minimalist runner, practising yogi and baker of all things whole-wheat, and sugar-free. Currently re-learning her ABC’s…the one that goes: A for algorithm, B for Bayesian, C for convolutional (neural network)….


Leave a Reply

Your email address will not be published. Required fields are marked *

Next post