Vijay Gopalakrishnan (Partner, Offering Leader for AI/Gen AI, Automation, BPM, and Strategy Deloitte India), Zahra Cader (ESG, Government & Public Services Leader Deloitte Sri Lanka and Maldives), and Nishan Mendis (Technology & Transformation Leader Deloitte Sri Lanka and Maldives) explore how Sri Lanka can unlock the potential of AI for the benefit of every citizen.
Digital public infrastructure and digital public goods integrated with AI offer Sri Lanka opportunities to improve governance, streamline public services, and support economic growth. Digital ID systems can expand the tax base, improve welfare programme efficiency, and reduce unnecessary expenditures. Linking databases and collaborating with the private sector can enable traceable transactions, lower business costs, and improve service delivery.
While AI applications, supported by digital infrastructure, can impact sectors such as agriculture, healthcare, education, and urban planning, there are challenges around privacy concerns, ethical risks, and biases in AI systems, which may deepen inequalities if not addressed.
In a recent episode of Deloitte: Decision Points, Vijay, Zahra and Nishan discussed overcoming the challenges and leveraging existing infrastructure and global examples for a functional and inclusive digital ecosystem that will leave no one behind.
How would you define digital public infrastructure?
Vijay: Infrastructure, in the digital sense, is akin to a highway, road, or railway network. It consists of digital systems that allow governments, businesses, and citizens to engage in digital payments and transactions and access goods and services, much like how physical transportation networks enable the movement of people and goods.
Vijay Gopalakrishnan – Partner, Offering Leader for AI/Gen AI, Automation, BPM, and Strategy Deloitte India
The government supports the development of digital public infrastructure but collaborates with private companies to build it. This infrastructure can then be leveraged for various analytics use cases as analytics requires data and a digital backbone to function effectively.
What is a digital public good?
Vijay: In the context of India, the digital public infrastructure consists of components like Aadhaar, a biometric-based national identity system, the Unified Payments Interface (UPI) for mobile payments, and DigiLocker for securely storing digital financial assets. These components are digital public goods, in essence. They are the tools which enable citizens to perform various digital services and transactions using the digital public infrastructure. Digital Public goods are built using open-source software with open standards to help use the best practices and be more inclusive without dependency on a specific technology.
India has made giant strides in digital public infrastructure, particularly in the last five years since the introduction of Aadhaar. What can you tell us about this?
Vijay: The IT services revolution in India laid the foundation for significant digital initiatives, including Aadhaar, the biometric-based unified identity database. The success of such initiatives was supported by the telecom revolution, with companies like Airtel and government backing, enabling necessary technology and network infrastructure. Aadhaar plays a crucial role in providing identity to all sections of society, including nomadic or disadvantaged groups which lack other forms of identification. It enables them, for example, to enter the banking system by using their unique Aadhaar identity to open a bank account easily.
What can Sri Lanka learn from India?
Nishan: Sri Lanka faces challenges with increasing government revenue. The chosen solution has been to raise tax rates, but this approach is impractical and unsustainable in the long term. Instead of increasing tax rates, Sri Lanka should focus on expanding the tax base, and digital ID can play a crucial role in this. By linking various systems like the land registry, immigration, and banking, a digital ID would allow the Inland Revenue to track financial activities more efficiently, helping identify who should pay taxes.
Additionally, digital payments and a cashless economy would make transactions traceable, further aiding tax identification. Regarding public expenditure, digital ID can reduce wasteful spending by ensuring welfare benefits reach the right people. For example, in India, Aadhaar saved significant amounts of welfare payments by eliminating fraud. Digital ID can also help prevent payments to deceased individuals and improve efficiency in welfare distribution.
Furthermore, digital infrastructure like e-KYC (Know Your Customer) can reduce business costs and improve productivity by lowering transaction costs, enhancing transparency, and streamlining processes in both public and private sectors.
Could you share examples of digital public infrastructure and goods enabling AI applications to serve citizens, including economically disadvantaged communities?
Vijay: Aadhaar, the biometric-based ID system, has helped underserved small-scale farmers access credit. Previously, these farmers struggled to obtain loans due to a lack of proper identification and financial documentation. By enabling them to open bank accounts and conduct financial transactions, Aadhaar has created a financial history for these farmers. This transaction data can now be used in predictive AI models for credit risk assessment, allowing farmers to access microloans.
What are the primary risks of AI adoption in government and social sectors? How can we ensure that doing so does not exacerbate inequalities in society?
Zahra: AI offers opportunities for positive change, but its impact depends on usage. Key risks, particularly for disadvantaged communities, involve privacy and ethics, as data from sectors like telecom, banking, and healthcare often gets aggregated without individuals knowing what happens to their data.
Zahra Cader – ESG, Government & Public Services Leader Deloitte Sri Lanka and Maldives
Additionally, AI systems can reflect biases present in past data, leading to unfair outcomes, such as over-policing communities or perpetuating gender disparities in hiring. Stronger regulations on data use and biases can address these issues. AI should serve as a decision support system. It should not be making decisions independently. Human intervention and oversight, a human-in-the-loop, are crucial to ensure ethical AI practices.
Another challenge in AI adoption is the lack of diverse digital data. Many existing data sets may not represent the entire community and may overlook the concerns of disadvantaged groups. The quality and diversity of data are crucial for ensuring that AI systems are inclusive and address the needs of all groups, including those who may otherwise be affected or excluded from the process.
Should Sri Lanka adopt something similar to what India has done? Or should we look at a homegrown solution to digital public infrastructure and goods?
Nishan: Culturally, the two countries share many similarities and challenges, particularly in data asymmetry. While there are differences between the two countries, such as population size, Sri Lanka can draw on the Indian approach, adapting it to fit its context.
India offers valuable lessons. India has evolved, made mistakes, and corrected them. Sri Lanka does not need to reinvent the wheel. I had the opportunity to talk with the chief architect of the Aadhaar project, who explained that many of the challenges they encountered are similar to those that Sri Lanka faces now. The solutions India applied can fit the Sri Lankan context.
Additionally, Sri Lanka can leapfrog certain stages of development. A specific example is how India is now evolving Aadhaar to allow citizens to share only part of their data, such as redacting birth dates to protect privacy. Sri Lanka can take advantage of these advancements, also avoiding the mistakes India would have made in the early stages of the digital identity system.
Moreover, Sri Lanka already has some foundational digital infrastructure, such as NIC databases, passport records, and biometric data, so it does not need to start from scratch. By leveraging existing data and experiences from India, Sri Lanka can adopt a hybrid approach to build its digital public infrastructure efficiently.
Vijay: Sri Lanka can leverage existing infrastructure to quickly build its digital public services, especially in areas where India had to start from scratch. Sri Lanka has biometric data for a significant portion of the population who have passports, so that is an advantage. Unlike India, which had to create a unique biometric identity database from the ground up, Sri Lanka can build on available resources, thus saving time and resources. That means Sri Lanka does not need to spend as much effort creating a new biometric identity system, allowing it to move more quickly with AI-driven public use cases. Additionally, Sri Lanka can take a phased approach, rolling out digital public infrastructure in smaller, incremental steps rather than waiting for a fully complete system. This approach, also adopted by India, helps deliver value early while progressively expanding the system.
What are some of the AI use cases that have been implemented globally and regionally with digital public infrastructure and goods?
Vijay: Other valuable data sources exist, particularly in the private sector, apart from the government data sources from digital public infrastructure and goods. For instance, telecom companies possess citizen data, including locational information, which can enhance AI models’ accuracy and efficiency. When creating AI solutions for government projects, it is best to integrate public and private data sources into a common database accessible to the AI system. This approach can be applied across various sectors, such as education, skills development, and social programmes, leveraging a combination of government and private sector data to improve outcomes.
In India, AI is used in various sectors to improve efficiency and outcomes. In the education space, facial recognition and expression-based AI tracks student engagement in the classroom, providing valuable feedback to teachers and enhancing teaching quality. In urban planning, AI-driven traffic signals in Bangalore manage traffic flow based on real-time actual traffic patterns. AI is aiding in finance by enabling credit modelling for rural farmers. Healthcare benefits from apps that e.g. track COVID-19 status and alert individuals at risk through locational analytics. In agriculture, satellite imagery and historical data-based AI guide farmers in improving crop yields through soil restoration techniques. For land property rights, Generative AI is helping speed up property title verification by scanning and summarizing documents. These AI applications have rapidly moved from proof of concept to practical use at scale in the private and public sectors.
Nishan: AI’s core value lies in its ability to process large volumes of data. With vast data sets, traditional decision-making becomes difficult, which is where AI and data analytics tools come in. These tools are essential for quickly making sense of the data, allowing for informed decision-making. Additionally, AI plays a role in building personalized services and improving service delivery and customer satisfaction.
How might AI-driven initiatives help countries achieve their Sustainable Development Goals?
Zahra: The SDGs for 2030 have 17 goals and 169 targets, but progress towards achieving them is currently behind schedule. Technology, especially AI, is essential in accelerating progress towards these goals. AI is already making significant contributions to healthcare, education, climate, and development. In healthcare, AI is helping with early disease detection, particularly cancer, through medical imaging and data algorithms, improving patient outcomes and optimizing resource allocation. AI is also aiding in predicting patient inflows, enabling more timely and efficient care.
In agriculture, AI is improving productivity and resource management, furthering the impact on SDGs by enabling precision farming, where specific fertilizers and pesticides are applied based on soil requirements and weather patterns, leading to resource optimization and contributing to sustainable practices. This approach supports broader SDG goals, particularly resource conservation and food security. In the climate sector, AI, combined with satellite imaging and ground sensors, helps predict emissions, monitor air quality, and help governments develop policies and regulations. These AI-driven solutions contribute to achieving SDGs by enhancing environmental sustainability and resource efficiency.
When Sri Lanka implements digital IDs, who should be at the table to leverage AI on the available data? Should the government be the sole stakeholder?
Nishan: The government is critical for establishing systems, particularly AI and digital infrastructure. It acts as a key stakeholder due to its responsibility for creating the right policy and regulatory environment. Strong leadership from the government is crucial, as it must make timely and sometimes difficult decisions that may not be popular with all parties involved. Effective leadership is a success factor for the rollout and execution of such initiatives.
Will there have to be an AI regulator or a government unit on AI? How does it work in India?
Vijay: India and Sri Lanka are formulating national AI strategies, but there is no need to wait for them to be fully in place before taking action. The government can drive AI initiatives, provide leadership, and set the regulatory environment. Without government support, such projects would not succeed. The private sector plays a crucial role in this process, with companies like Gupshup and RS Software contributing to digital public infrastructure development in India because the government created an enabling environment. Strong collaboration between government leadership and private sector expertise is crucial for success.
Success in India was due to the government’s proactive role in seeking leadership and advice from experts while also giving them the flexibility to drive the project. Equally important was the government’s ongoing support over the years, regardless of leadership changes. This continuity is crucial, as any disruption could cause stakeholders to lose confidence and withdraw their support.
If the government is the central player, who could the other stakeholders around the table be?
Nishan: Private sector involvement is crucial for the success of digital public infrastructure and goods, as demonstrated by Aadhaar and UPI initiatives. The Indian government laid the foundational infrastructure, but the real success came from opening it up to private players, including startups, who developed diverse platforms and applications. The key to multiplying the benefits of such systems is creating an ecosystem where various players can contribute rather than relying on siloed government systems. This collaborative approach ensures broader adoption and innovation, driving national progress.
Zahra: Bridging the digital divide is a crucial component of successful AI adoption, together with its technical and policy aspects, ensuring that all communities have access to digital infrastructure and skills, particularly at the grassroots level. Programmes like embedding AI into early education are crucial for long-term skill development. Reskilling programmes for the existing workforce are also necessary as AI may replace some routine jobs, and people need help adjusting to new roles in an AI-driven economy. These social considerations are essential for inclusive growth and ensuring that AI benefits everyone.
How do you deal with the fear that AI might lead to job losses? How do you address concerns like that because what you do not want is a pushback?
Nishan: The concern about AI replacing jobs is valid, but historically, technological advancements, like previous industrial revolutions, have led to the creation of new jobs rather than widespread unemployment. AI is automating routine tasks, freeing time for humans to focus on creative and decision-making roles. While AI will replace routine jobs, it will generate new types of work, particularly in areas requiring human creativity and problem-solving. The key to this transition is reskilling and upskilling the workforce, ensuring people are ready for the evolving job market.
How do you allay fears about the impact of AI on ordinary people? What did India do?
Vijay: The fear of job loss due to technology is not new. In India, decades ago, there were protests about computers taking away jobs, yet technology has ultimately transformed society and created more jobs. The landscape has changed, and instead of losing employment, more people now have jobs.
The growth in sectors like IT has led to broader economic benefits, especially in cities like Bangalore. While IT jobs have grown, adjacent industries, such as hospitality, retail, and transportation, have flourished and created even more jobs. Effective messaging, exemplified by public campaigns and leadership such as the Prime Ministerial addresses, can alleviate fears about job loss from technology. Historical examples show that the fear of job loss due to new technology is unfounded, and in fact, technology has created more opportunities, not fewer.
How do you reskill people to deal with this new reality at scale?
Vijay: Even in the worst-case scenario where jobs like taxi drivers are affected by automation like driverless cars, such changes will not happen overnight. Implementing technologies like driverless cars in India will take years. Moreover, the government would not allow large segments of society to lose their livelihoods, as this would lead to significant social issues. It should introduce changes in phases to give time for reskilling. Academia, universities, and the private sector all play roles in this process, helping to train the workforce for new jobs and to ensure a smooth transition.
Do you think inequalities worsen at the lower end of the income pyramid, as wealthier individuals can acquire skills more easily?
Vijay: India has made significant strides in literacy, education, and accessibility. For example, when I was growing up in India, poverty was visible. Not so now. Even taxi drivers who once lacked English skills can now read simple messages and send texts. Government initiatives like using chatbots and digital platforms are helping people, especially in remote areas, access information on career opportunities and skill development in their native languages.
A key initiative is to enhance English fluency among the workforce by 2030. The rapid evolution of telecom, through initiatives like Reliance Jio, has made smartphones and data services more accessible, enabling greater participation in digital payments and access to online resources. With the availability of language models in various Indian languages, even non-English speakers can now take advantage of these technologies, making their adoption much easier and faster than anticipated.
What are the low-hanging fruits for others in society? What about the private sector companies?
Nishan: The examples from India highlight how digital public goods and infrastructure can increase efficiency, reduce business costs, and improve access to goods and services. For instance, the digital KYC process lowers transaction costs and enhances public service delivery. One of the key benefits of these digital systems is improving transparency and reducing corruption in public sector interactions by eliminating face-to-face meetings, which often lead to bribery or delays.
Nishan Mendis – Technology & Transformation Leader Deloitte Sri Lanka and Maldives
Additionally, foundational digital platforms can support various applications, such as offering online education in multiple languages. One example is the Barshini initiative, which provides translations in nearly 20 languages, making services more accessible to more people. These digital advancements help streamline processes and benefit society by providing more efficient, transparent, and accessible services.
Vijay: The unique biometric ID system helps ensure inclusivity, by e.g. helping people from disadvantaged communities to create bank accounts using the ID and to enter the banking system. It makes it harder for individuals to evade taxes, as the ID is linked to all personal data, creating a traceable record. This system increases accountability and helps ensure that services and benefits reach disadvantaged communities since payments are made to the bank account of a person which is linked to his or her unique biometric ID.
Has it led to greater income tax compliance or tax compliance in general?
Vijay: Digital technology, like digital payments and Aadhaar-linked systems, has significantly increased income tax compliance and revenue collection, making it harder for individuals to evade taxes. By logging payments and linking them to unique identifiers, the government can better track financial transactions and reduce fraud. Additionally, verification has become faster and cheaper, leading to cost savings. This increased revenue allows the government to fund subsidies and development projects for underserved communities. Furthermore, biometric Aadhaar-linked bank accounts ensure that subsidies reach the correct recipients. The digital public infrastructure, similar to a road or railway network, is being leveraged by both public and private sectors to develop innovative solutions, enhancing overall efficiency and accessibility.
The example of BharatPe in India illustrates how a component of the digital public infrastructure, like the Unified Payments Interface, enables interoperability across various payment apps, making transactions seamless. BharatPe built a business model by offering a single QR code for payments, regardless of the mobile payment app used, and later expanded into loans to small retailers.
By the way, on a separate topic which we discussed earlier, AI applications in the public and private sectors have well-established techniques to prevent biases, provided the AI models are trained on the appropriate data and have proper guardrails put in place.
Nishan: The Unified Payments Interface has reached grassroots levels across India, benefiting even small vendors selling King Coconuts or tea at local wayside stalls. These vendors now use QR codes for payments and transactions through banking channels. The Unified Payments Interface has increased efficiency and economic benefits, as users can digitally track all their transactions. The vendors, not the government, drove the adoption of this system because they saw its value to their livelihoods.
Vijay: The adoption of mobile payments by small vendors in India accelerated due to external events like COVID. Before this, vendors had to manage cash, risking theft and the inconvenience of depositing it in a bank. With mobile payments, they no longer face these challenges. While the initial triggers were related to economic disruptions, vendors embraced the technology, leading to faster adoption and positive outcomes, such as greater security and efficiency.
Any closing thoughts?
Zahra: AI has significant potential to benefit society. It requires collaboration among all stakeholders. By addressing the associated risks and putting the appropriate structures in place, Sri Lanka can harness AI for the greater good.
Nishan: Success depends on collaboration between the government, academia, and the private sector. A key focus should be on reskilling and upskilling the workforce to ensure people can effectively use and adapt to new technologies. Academia plays a crucial role in this process.
Vijay: South Asia, particularly India and Sri Lanka, has the potential to lead in AI innovation. India has successful use cases of AI benefiting local and global populations. With its high literacy and strong social and economic foundation, Sri Lanka is well-positioned to succeed in AI, provided all stakeholders collaborate and move forward together.