Need to Build Narrowly Focussed and Clearly Defined AI Use Cases

Perspectives shared by Shri Sunil Kumar, DDG, NIC

The journey of eGovernance is clearly transcending from automation to analytics. The next logical milestone will be to build systemic intelligence by exploiting emerging technologies like Artificial Intelligence, Machine Learning, Deep Learning, Natural Language Processing, Virtual Reality etc. How more and more complex tasks are automated with built-in intelligence will determine how much effectiveness and efficiency is achieved in serving the citizens.

AI and other related fields can have a very broad spectrum of applications, ranging from narrowly defined tasks to a broader and deeper usage. To begin with, it’s the narrow end that will get more attention and efforts.

– Shri Sunil Kumar, DDG, NIC

That’s only logical, it’s very important to build large number of use cases around narrow applications of AI, as it will take time for right conditions to evolve for a broader and deeper adoption of the newer technologies. The perspective one gets from these use cases can help scale them up, include adjoining processes within domain, broad-base the usage to multiple domains, and mainstream AI in governance.

Shri Sunil Kumar, DDG, NIC

That’s the philosophy with which we have decided upon using ‘chatbots for frequently asked questions (FAQs)’ by harnessing natural language processing as a use case for AI. Many citizens are interacting with government, either to enquire about their service requests or register their grievances or seek relevant information. They currently go to a physical centre or connect over the phone or web, all requiring human intervention.

Now, though the number of interactions is huge, there are fewer and manageable number of ‘type of issues’ that are covered during the interactions. The process of identifying the generic question ‘types’ and their standard answers, processing query in natural language and providing answers also in the natural language can be automated using appropriate technologies. A machine learning engine can work at the back end to decipher and match the query, a natural language interface can work at the front end, both to receive the query and address the same with an appropriate response.

Read the Complete Best of Tech Report

The use case will have very strong implications as the citizens can now interact and get response in their own natural language. Their multi-purpose mobile phone in its simplest form can be used for voice query. The initiative will enhance both the effectiveness and efficiency of service provisioning, the extent depending upon the degree to which the application is integrated with the back-end service provisioning engine.

Coming Soon

We are exploring this application in Reproductive and Child Health Scheme (RCH), Central Government Health Scheme (CGHS), and Online and Registration System (ORS) of eHospital. The same can be explored in many other domains of governance.

Most Popular

To Top