National Informatics Centre (NIC) is betting big on Artificial Intelligence (AI) for smooth implementation of several e-governance projects of the Centre and thorough transparency in execution of welfare schemes.
AI is one of the emerging technologies which is benefitting not only the government agencies but also the citizens. AI is helping in re-imagining the way various services such as financial services fraud detection, online customer support interactions and retail purchase predictions, etc., are being carried out.
NIC’s Centre of Excellence in Artificial Intelligence (COE-AI@NIC), unveiled by Minister of Electronics & IT, Law &Justice, Ravi Shankar Prasad last year, is equipped with an ultra modern lab with supercomputing facility, which has facilitated AI Development Platform as a Service to twelve NIC State Units and three NIC Central Project teams.
NIC has prepared a handbook to build in-house expertise in augmenting intelligence in design, development and usage of AI technologies to improve citizen services delivery.
The Centre of Excellence in AI of NIC has been facilitating model development in computer vision like image and video analytics, text analytics and cognitive search, conversational AI in the form of chatbots and voice bots, machine learning and trend forecasting.
NIC also helps citizens get the benefits from the scheme, reducing the time required in government workflow processes, by enhancing transparency in the process cycle through intelligence augmentation in process steps.
In SBM Urban, the citizens apply for these schemes online directly through the platform or at Citizen Service Centres and submit documents like identity credentials, bank account details, applicant photo which helps in validating their identity and establish their eligibility. The beneficiaries upload photos of progress of work to get installments from these portals through Direct Benefit Transfer (DBT). For establishing right utilisation of funds under these schemes, physical verification checks are conducted by officials before transfer of funds to the beneficiary’s bank account.
In SBM (Urban) which facilitates people from the lower strata of the society to build toilets under Indian Household Latrine (IHHL) scheme, beneficiaries upload their photo at application time and later upload photo with the constructed toilet for receiving the final installment in DBT to their bank account. In SBM Urban, there were over a crore applicants and more than half a crore approved beneficiaries who uploaded geo-tagged constructed toilet photos through online platforms. Earlier, there used to be a delay in fund transfer to beneficiaries account. This was mainly because the photo may get rejected by the verifier and the beneficiary had to again upload the photo and wait for the final nod. Helpdesk used to receive huge number of complaints pertaining to this issue.
NIC started using AI to address these challenges of SBM. AI was used to bring down the workflow cycle for completion of the procedure. If the system can detect there is no beneficiary in the image, or toilet seat is not visible then the beneficiary can check the status with a mobile app or be alerted by SMS to reload correct photo without wasting time and the citizen does not have to wait for the photo to be rejected before reloading the correct photo and waiting in queue for approval.
Another such example where the benefits of AI has been explored by NIC is the facial recognition based group attendance for traineesfor Skill Development Program, under the Department of Technical Education, Training & Skill Development, Government of West Bengal. Under Utkarsh Bangla program, every year six lakh trainees are trained in skill development programs. The time for imparting training is being calculated through the biometric attendance system for the trainers and trainees. Payments of training partners are calculated based on the total number of training hours and count of trainees who have attended the classes. Every year, the government spends around Rs 200 crore for organising these training programs.
Existing fingerprint based biometric attendance system is computation intensive and could be circumvented through various means including use of prosthetic fingerprints. Also, using same fingerprint sensor by hundreds of trainees daily may cause health hazard and a challenge to maintain personal hygiene. Using facial recognition techniques, with the help of a mobile based app, the face of trainees in a particular batch for a skill centre are registered as a one-time activity and then used in detecting the attendance on a daily basis. Multi factor authentication like geo-fencing of the Training Centre in the image, timestamping are built into the system and while marking attendance, the system prompts a particular trainee to be placed in Left or Right of the image, to make sure that there is no video replay attack of the group photo.
NIC Director General Neeta Verma said, “Organisations and businesses today are realising only a fraction of their true AI potential. Leaders, be it from the government or the private sector, have explored the benefits of plugging AI and other tech tools into existing workflows, focusing on automation and execution to improve existing services.” But simply focusing on using AI to make their existing activities run faster and economic might limit its impact. With emerging technologies taking the center stage, leaders need to leverage the potential of AI systems to transform not just how organisations perform, but also what they actually do. AI is certainly becoming an agent of change across the organisation, she said.
Besides, AI is also being used in the power sector where it plays a key role in analysing reports, graphs, statistics for generation, transmission and distribution. NIC has developed National Power Portal (NPP), a centralised platform for collation and dissemination of Indian power sector information. NPP was integrated with associated systems of Central Electricity Authority (CEA), Power Finance Corporation (PFC), Rural Electrification Corporation (REC) and other major utilities to serve as a single authentic source of power sector information to apex bodies. It facilitates online data capture, input (daily, monthly, annually) from generation, transmission and distribution utilities in country. It also disseminates Power Sector Information (operational, capacity, demand, supply, consumption) through various analysed reports, graphs, statistics for generation, transmission and distribution at central, state and private sector level. The dashboard has been designed and developed to disseminate information about the sector through GIS-enabled navigation and visualisation chart windows on capacity, generation, transmission, distribution at national, state, DIS-COM, town, feeder level and scheme-based funding to states.
With the increasing decentralisation and digitalisation of the power grid, it is becoming challenging to manage the large number of grid participants and keep the grid in balance. AI can help in detecting anomalies in generation, consumption, or transmission in near real time, and then develop suitable solutions. As a first prototype, deep learning is being attempted to see the power outage trends and make predictions for the next subsequent months from the attributes captured. An accurate prediction of duration of power failures for the upcoming months can help the authorities to minimise the possible risks which may cause such outages.
