Explained | Digital Public Infrastructure & AI

India has pitched for global adoption of Digital Public Infrastructure (DPI) to promote productivity and achieve inclusive and sustainable growth. The ‘Report of India’s G20 Task Force on Digital Public Infrastructure’ by India’s G20 Task Force was released in New Delhi.

Many countries are considering how to develop their national digital infrastructure to accelerate economic progress through drastic improvement in provisioning public services, and foster trust between people and institutions by improving transparency and reducing distance. 

The report will play a key role in defining the future course of the DPI approach and actions for implementation around the globe, particularly in the Global South.

Digital Public Infrastructure (DPI)

• Digital Public Infrastructure (DPI) refers to platforms such as identification, payment and data exchange systems that help countries deliver vital services to their people. 

• The DPI is an approach designed to solve multi-faceted complex societal problems that ultimately require both public and private participation and innovation.

• It allows nations to fast-track their economic development in an inclusive and competitive manner, leveraging cutting-edge technology, good governance and markets to drive outcomes such as health and financial inclusion, human development, access to jobs and employment, and improved public services.

• DPI drives innovation, competition and inclusion, and is governed by enabling rules and relies on a technology layer, governance framework and a multistakeholder ecosystem.

• Verifiable identity, electronic KYC, interoperable fast payments networks, and consented data sharing frameworks that offer open, interoperable solutions to societal problems at the population scale are strong

examples of DPI. 

• India has built DPI across several sectors including foundational identity, payments, consented data sharing, education, health, commerce, etc.

• Further, with the appropriate application of emerging technologies, including artificial intelligence (AI), the impact of DPI can easily be amplified further. 

• It is also important to continuously focus on building new DPI and increasing adoption by incorporating feedback and encouraging innovation as and when possible.

Categories of DPI

DPIs can be categorised into several groups which fall within and across multiple sectors.

1) Verifiable Identity

Digital identity is a necessary prerequisite in order to be able to use a digital service. Identity makes it possible for those features and benefits intended to be provided exclusively to a given user to be provided to her and not anyone else. 

Verifiable identity should be interoperable by design so that it is capable of being used across a range of different systems to digitally authenticate users in a variety of different ways. Once deployed at national scale they can form the basis of such initiatives as bank account opening, mobile SIM registration, or government-to-person payments.

2) Interoperable Payments

Instant and interoperable payment systems are another essential DPI building block. These systems establish open, interoperable digital infrastructure that make it possible to quickly and seamlessly transfer money from one account to another (across any kind of device, any institution, any kind of account, and any type of recipient) while at the same time mitigating the risks of exposing potentially sensitive financial data.  

In the long run, inclusion to digital payments systems drives up commerce at scale and helps individuals build granular financial history for access to services like creditor insurance.

India’s Unified Payments Interface (UPI) is a strong example of digital public infrastructure in the person-to-person (P2P) / person-to-merchant (P2M) payments space. It leverages the ubiquity of smartphones and the convenience of mobile internet-based solutions to offer unprecedented convenience, and drives inclusion by offering payment modes on feature phones, smartphones, via biometrics (without phones) and via voice instruction. 

India’s Bharat Bill Payments System, Aadhaar Payment Bridge for government-to-person (G2P) Direct Benefit transfers, FASTag toll collection system, and Aadhaar Enabled Payments System for biometric cash deposit & withdrawal are also strong examples of interoperable payments building blocks solving for bill payments, G2P payments, toll payments, and cash deposits and withdrawals respectively.

3) Open Networks for e-Commerce

Today goods and services are offered online through massive vertically integrated platforms that have resulted in consolidation of buyers and sellers in ways that offer the gatekeepers extraordinary power and influence over commerce. By breaking the commercial transaction into its constituent elements — inventory search, order booking, fulfilment and post-fulfilment resolution — DPI solutions can offer more equitable alternatives.

India’s Open Network for Digital Commerce (ONDC) is a good example of an open network DPI. It is an interoperable, inclusive and open standard based ecommerce marketplace that enables effortless product visibility and efficient sale transactions between the buyer and seller applications. The ONDC improves efficiency in the tasks like product listing, inventory management, and order fulfilment processes, empowering small businesses to enhance their presence and effectiveness in e-commerce space.

4) Consented Data Sharing

DPI can be designed to liberate data from the silos in which they currently reside so that the data principals to whom they pertain can extract greater value out of them than is currently possible. This is manifested through consented data transfer frameworks such as India’s Data Empowerment and Protection Architecture (DEPA), that have a range of applications across different sectors. 

India’s DEPA empowers individuals by giving them ownership over their data (rather than being controlled by the service providers who collect this data) and allowing them to share it across providers to enable access to tailored products and services.

5) Open APIs for Govt Services

While the digitisation of government services or the aggregation of government services in a common portal or platform is not itself considered a DPI, the publication of open Application Programming Interfaces (APIs) for government services is considered a DPI. These could include open APIs for tax filing, business registration, identity authentication, application for government schemes, application for various licenses, etc. 

For instance, India’s ‘API Setu’ is an example, as well as the open APIs published by Goods and Services Tax Network (GSTN) which allow a number of third-party apps to design unique and inclusive user experiences for filing taxes.

6) Registries

Registries that are considered DPI should be machine readable and digitally signed records accessible to private players via open APIs that allow discoverability and verification for any nouns in a digital economy.

For instance, farmer registries, health facility registries, teacher registries, or vehicle registries are common examples.

7) Electronic Signatures (eSign)

Electronic signatures transform document handling and signatures by creating a paperless environment, removing the complexities of physical copies and courier delays. It streamlines processes, saves resources, and promotes eco-friendliness. Real-time document signing eliminates waiting periods, speeding up processes and reducing costs. With eSign, businesses can make faster decisions and reduce operational delays. Aadhaar eSign has revolutionised banking in India with numerous applications and benefits.

Artificial Intelligence (AI) and DPI

• In the rapidly evolving technological environment, AI stands out as a new tool offering vast opportunities and the ability to redefine the future. Integrating this cutting-edge AI technology into the robust and

responsible Digital Public Infrastructure presents an opportunity to amplify its capabilities and efficiency even further. 

• The seamless fusion of DPI with AI propels us into a new world of ‘Digital Public Intelligence’, representing a transformative leap forward.

• The function of AI is to interpret intricate and complex data, foresee requirements, and offer personalised solutions customisable according to individual preferences and situations. 

• Therefore, AI can facilitate DPI’s ability to evolve continuously in real-time through various ways.

• Nations like India, characterised by a multitude of local languages, consistently encounter data availability challenges. In such contexts, AI can help overcome the challenge of limited data availability to a large extent in Indian languages by enabling language localisation.

• Natural Language Processing (NLP) models can be trained on diverse datasets to understand and process Indian languages, making DPI more inclusive and accessible to all citizens. 

• Bhashini is one such example of NLP that has been made to understand, process, and produce text in a variety of Indian languages.

• DPI can use AI to enhance service delivery across various sectors such as health, education, agriculture, and finance. AI algorithms can analyse large datasets to provide insights and recommendations, improving

decision-making processes and resource allocation. It can help to streamline processes, automate routine tasks, and provide personalised services based on data analysis.

• AI can also bolster cybersecurity measures within DPI by continuously monitoring for threats and anomalies. AI-driven systems can detect and respond to security breaches more rapidly, reducing the risk of data

breaches or cyber-attacks.

• Further, AI can foster innovation in DPI. Open-source software, AI models, standards, or content that are publicly available can be replicated, modified, and shared freely. This encourages innovation and scalability in DPI by promoting the entry of private players in the technology industry.

• Considering the potential risks tied to emerging technologies like AI, governments need to nurture the growth of a digital environment conducive to trustworthy AI. This ecosystem encompasses various digital technologies and infrastructure, notably DPI. 

• DPI plays a pivotal role in upholding ethical and responsible AI standards, emphasizing fairness, transparency, and accountability in AI-driven decision-making processes. 

• By integrating principles like safeguarding privacy, mitigating bias, and ensuring algorithmic transparency, DPI can instill trust and assurance among citizens regarding AI-powered services.

• Given such challenges associated with the new technologies such as AI, governments should foster the development of, and access to, a digital ecosystem for trustworthy AI. Such an ecosystem includes in particular digital technologies and infrastructure such as DPI, and mechanisms for sharing AI knowledge, as appropriate. 

• By incorporating principles such as privacy preservation, bias mitigation, and algorithmic transparency, DPI can build trust and confidence among citizens in AI-enabled services.

• A DPI approach to craft the growth of AI would involve publication of open datasets via open API to train AI/machine language models, creation of reusable AI toolkits (such as data tagging and curation tools) and publication of open models to drive market innovation in AI in service of all individuals.