Artificial Intelligence (AI) has made tremendous advancements in number of domains such as healthcare, telecom, banking etc. However, its deployment in the e-government applications is still hindered. Experts cite that the scale and impact at which AI can help solve challenges pertaining to public and social good is tremendous and inevitable.
- Artificial Intelligence(AI) can be defined as the ability of computer to imitate human intelligence while improving its own performance.
- Machine Learning (ML) is the ability of an algorithm to learn from prior data in order to produce a smart behavior that it has never faced before.
- Deep Learning (DL) can be defined as mapping function that maps raw input data (X-ray image) to the desired output (diagnosis) by minimizing a loss function.
Challenges inherent to adopting AI techniques in e-governance domain
- Given the recent and rapid advances in deep learning domain, it is becoming more difficult to find experts of this technology.
- Deep learning application development is largely unsystematic and based on optimizing large set of parameters. It doesn’t follows conventional development life cycle of meeting defined set of functional and non-functional requirements.
- AI and deep learning applications require strong policies on data security and privacy.
“Advances in Artificial Intelligence (AI) present an opportunity to build better tools and solution to help solve some of world’s most pressing challenges, and deliver positive social impact in accordance with the priorities outlined in the United Nations’ 17 Sustainable Development Goals (SDGs)”
Management of Government Information Resources
Huge amounts of data is being generated every second from multiple sources. The data is of all kinds – homogenous, heterogenous, structured, unstructured, videos, audio, text etc. Proper management of information resources is the need of hour and include collecting, storage and processing of end-user data.
Framework for Centralized Management of e-Government information resources
Below framework is cited from an IEEE reference. The framework mainly consists of 4 components:
- Government Collective Office Network Unit – Responsible for implementing and ensuring correctness of e-government services and policies.
- Big Data Services Center Unit – Responsible for all processes and policies regarding Big Data (collecting, storing, processing and transmitting). This unit also plays a critical role in privacy and security of citizen and government data.
- Social Public and Research Unit – Provides e-Services for citizens and research organizations. It also includes research agency concerned with advancing the current state of e-government system.
- Intelligent Archives Unit – Responsible for digitizing paper documents and provinding smart & personalized services to other units that require accessing and consuming digital data.
These components utilize advances in technology such as Internet of Things, AI, storage utilities etc.
Automating e-Government with AI
Despite the availability of plethora of resource and applications, data is not being utilized to advance e-government services in data-driven approach. Utilizing deep learning can significantly improve the current state of services in domains such as computer vision and natural language processing (NLP). Some of the model examples that may benefit e-governance projects are – Hand Written Letters Recognition, Hand Written Digits Recognition and Sentiment Analysis.
Sentiment analysis may ideally be integrated in e-government services to analyze the overall efficiency of these services leading to improvement in service quality and addressing citizen’s needs.
The many AI techniques such as tagging, part of speech recognition, classification, sentiment analysis, tokenization and entity extraction may also be used in automated archive and indexing of government and citizen data.
Some other use-case are listed under:
Public Welfare: Public data feeds, public reporting and surveillance data could potentially be used to identify individuals or population that is in danger of homelessness or with tendency to take paths to undesirable personal or social outcomes.
Education: Social feeds, school records and well constructed social-network informed behavioral models can be used to detect school populations that are at risk of falling below a certain grade level.
Public Health: Health record data can be mined to develop predictive models to detect high-risk population and predict infectious disease spread. Application may range from diagnostics, finding potential hot spots, monitoring nutrition levels and preventing lifestyle diseases.
Public safety and transportation: Good quality data many be mined for effective deployment of first responders, understand mobility patterns of people, identify gaps in transit concerning citizen needs etc.
Disaster Response: AI and computer vision methods applied to satellite and drone images may aid in preparation effort for natural disasters such as cyclone and wildfires.