“Machine intelligence is the last invention that humanity will ever need to make.”

– Nick Bostrom, Swedish philosopher

Introduction

The World Bank (WB) defines GovTech as the “whole-of-government approach to public sector modernisation and promotes simple, efficient, and transparent government with the citizen at the centre of reforms. 1 ” It is the next stage to e-governance and digital governance. Like FinTech brings new technologies for banks, MedTech brings new technologies to hospitals, and EdTech brings new technologies to universities, GovTech brings technology to the government sectors to make it efficient and responsive. According to the WB, GovTech emphasises three aspects of public sector modernisation: 1

It is a platform approach to government service delivery (Figure 1). For exam- ple, Microsoft relies on several third parties to develop software on its Windows platform. Similarly, Twitter and Facebook rely on third parties for value-added services using their platform. Amazon cannot manufacture everything and must rely on multiple third-party manufacturers and traders who sell on its platform. By giving third parties an application programming interface (API), these plat- forms create value for themselves, third parties, consumers and other stakeholders. Thus, the platform approach co-creates value for the entire ecosystem. Bringing the same approach to the government means allowing third parties to connect to the government’s open platform. These third parties could be startups, small and medium enterprises (SMEs) or any other entities, which unlock value for citizens through their solutions. They connect with the GovTech platform using appli- cation programming interfaces (APIs) and open data and propose monetisation proposals to the government. 2

Figure 1: The concept of GovTech from Forbes 3 , NACo 4

In the GovTech ecosystem, private-sector startups and innovative SMEs deliver technology products and services to public-sector clients, often using new and emerging technologies. 5 Many of these startups work on challenges in public policy areas or problems that the government is looking to solve. Here GovTech works as incubators for firms that provide solutions to the government. The priorities of the GovTech ecosystem include improved efficiency and greater accountability in the public sector and its interactions with citizens. They also need to build trust across the diverse stakeholders in the ecosystem to develop a thriving GovTech industry to serve the domestic public sector and contribute to national economic growth. The purpose is to bring improved and innovative public services. Take Indian Railways, for example. Its platform approach has spawned a whole ecosystem of value-added apps on iOS and Android, providing new information to citizens. The government may launch a similar platform to open the entire GovTech space for innovation and co-creation of value. The current legacy ecosystems are flawed, require critical capacity enhancement, and are infested with cybersecurity issues. With the COVID-19 pandemic, governments became aware of these flaws and embarked firmly on the path of digital transformation.

What do they gain? We must understand that governments are limited in inno- vation and are slow to respond. They also do not have access to quality and skilled resources for providing innovative services to citizens. Therefore, for a government to solve all its woes is next to impossible. They hire expensive consultants to solve their problems. In the GovTech approach, the government can do away with these costly contracts by obtaining innovative technology products and services from startups and SMEs.

What is Required to Implement GovTech?

GovTech vision is implemented with citizens at the centre, and all government services revolve around them. The platform requires seamless internet connectiv- ity, citizens’ digital ID and a mobile payment interface (Figure 2). All administra- tive departments are integrated, the data is stored on the cloud and citizens have a two-way engagement with the platform. Startups and SMEs can use analytics to drive actionable insights and provide innovative citizen services.

Governments are implementing GovTech in various ways. For example, as part of the GovTech program of Poland, its Ministry of Entrepreneurship and Technology is developing a search engine drawing from Business Intelligence ideas. The intelligence embedded in the system allows concerned stakeholders to search databases and swiftly share findings and knowledge. Similarly, as part of CivTech Scotland, the National Health Service Scotland has partnered with Lumera Health in the UK to deploy an online booking framework for patients seeking services in the outpatient department. It has the potential to reduce time lost in being unable to honour their appointments. Incubatees (Oxon Tech, Greeve Systems and Brandwith) of the GovTech Catalyst program in the UK have developed a product for the Mid and West Wales Fire and Rescue Service. The product seeks to make available a building schematic that pins the location of firefighters and highlights motion information that is not usual, all in real-time. MIX Ontario is the GovTech incubator in Canada. Its incubatee– Snow Angels Canada – has developed an online platform for the City of Barrie, connecting volunteers to residents who require shovelling snow. Several other examples exist the world over.

Figure 2: GovTech elements from OECD 6 , Digital Future Society 7 , NACo, World Bank Group 8

From e-governance to Digital Governance to GovTech

In 1980, with the computerisation of Indian Railways, India started on the path of modernisation. Fast forward 40 years from then, and we are among the leaders in GovTech. The concept of e-governance was adopted by India in the early 2000s and began with developing websites that supply basic information to citizens. The agenda was purely information transmission, but transparency – the primary motive of e-governance – was hardly important. Indian Railways began web bookings in 2003, and it was in the same year that the State Bank of India – the country’s largest bank – launched e-banking services. The services were basic, though. With the increased pressure from the private sector regarding fast and transparent services, the Government of India (GoI) is geared towards better delivery of services to its citizens. Several government services were made available in the following decade through the Internet. The government also started mod- ernising IT infrastructure with its impressive state-wide area network (SWAN) project. Optical fibre cables started being laid throughout the country. The gov- ernment also focused on modernising its public services, and citizens could apply for certificates online in most states of the country.

In 2012, the government launched Aadhaar – a unique identity for Indian citizens as the basis for all services. It, however, gained mass after 2014 with a new government taking over the reins. It brought the entire country under Aadhaar – the digital identity of Indian citizens. Bank accounts were mapped with the Aadhaar and PAN card, in an attempt to control tax evasion, misuse of money for foreign exchange violations and terror funding. In 2014, the government started its Digital India programme. It also launched its vision for Make in India and Smart Cities Mission as part of the Digital India programme. The demonetisation of Indian Currency (Rs 500 and Rs 2000 Notes) in November 2016 forced India to launch a united payments interface (UPI) that allowed citizens to conduct cashless transactions. Today, of course, Digital India is more than a dream. UPI was built on an open architecture that facilitated the entry of private players in digital for payment processing. The government also made substantial strides in procurement with its GeM (Government eProcurement) portal that allowed SMEs and MSMEs to participate in tenders.

The GoI’s move towards the Startup India programme is another feather in its cap, which brought significant changes in the way citizens look towards society. India wishes to move towards becoming an entrepreneurial economy where peo- ple do not depend on the government for jobs. China came out of poverty when Alibaba’s marketplace allowed its small manufacturing industry to find a customer base not only within China but throughout the world. With a solid push from the government, several incubators and accelerators mushroomed across the country. With Digital India, all government departments embarked on a digital agenda. People could pay electricity bills and property tax online, book a gas cylinder online, check land records, get passports and virtually interact with the govern- ment too.

But there were some lacunae. While the central government was trying to build architecture at the Union level, the states were doing the same at their level. As a result, several different platforms were developed without interoperability. These systems would not talk to each other. Therefore, despite the provisions of digital governance, a citizen did not have a seamless experience throughout the country. When people commonly migrate to different parts of the country for a job or relocation, they experience different IT services across different states. Not just between states and the Centre, people experienced differences between departments within the same state or at the Centre. A citizen repeatedly provides the same data to several departments for similar services. It is time to stitch these services together and provide a seamless governance experience to every citizen.

Where does India Stand?

The WB Group developed the GovTech Index to determine a country’s position on GovTech. It constructed the GovTech Maturity Index, mostly on the GovTech dataset, including indicators for 198 economies across the world. This index also indicates the transition of governments from conventional e-government to more recent digital government and further to the futuristic GovTech. A GTMI index of 0.75 and above signifies that a country is a GovTech leader and demonstrates advanced solutions and best practices in key areas. Out of 80 GovTech initiatives worldwide, 43 economies have good practices. 2

India’s GovTech index is above 0.75, along with 42 other countries. Thus, the GTMI places India among the leaders in government digital transformation. The question is, however, if India is ready for it. The WB Group’s vision of GovTech manifests a holistic view of digital governance and includes four areas of focus:

Supporting core government systems : It refers to modernising and integrating government systems that touch citizens directly, such as tax administration, public procurement, public financial management, public investment management and HR management. The approach includes developing an all-pervasive transforma- tion strategy as well as key concepts to facilitate the use of digital infrastructure, including platforms and data that are secure and allow for smooth exchange from one to another. India has several different portals which still do not talk to each other. For example, its tax portal is online and seamless and citizens can obtain tax refunds within a matter of two to three months. The GST system is entirely online and has overcome its teething problems. Similarly, public procurement happens through the Government E-Procurement portal (GeM). With the development of NSDL and CDSL – national share depositories incorporated by the markets regulator Securities and Exchange Board of India (SEBI) – public investment in share markets has gone digital and become more efficient. The government has also taken long strides in skill development through the National Skills Development Council or NSDC. Interoperability is an issue, and with a contained digital transformation strategy, these will improve further.

Enhancing public service delivery : GovTech supports “the design of human-centred online services that are simple, transparent and universally accessible”. 1 Such ser- vices should be accessible using low-cost devices and options, such as cell phones and free apps that can reach all intended beneficiaries without linguistic barriers. The GoI emphasises services that are designed to be inclusive and user-friendly. Most services are multilingual and accessible to people with disabilities.

Mainstreaming citizen engagement : The emphasis of this focus area is to ensure that citizens can easily provide feedback and launch complaints to seek solu- tions that are difficult to achieve otherwise. The enforcement and monitoring mechanisms should be built to advance a government’s efforts to achieve greater transparency. As of now, the government does not have its solutions, but it does respond to citizens on global social media platforms. The GoI is active on Twitter and responds to queries from citizens. Moreover, the government is working on apps that allow citizens who cannot use keypads to provide voice- based complaints or feedback.

Fostering GovTech enablers : The World Development Report 2016 – Digital Dividends 9 identifies three primary analogue enablers of GovTech. These are effective regulations (an appropriate and conducive legal and regulatory regime), improved technical skills (digital skills in the public sector) and accountable institutions (strong enabling and safeguarding institutions and the environment that fosters innovation in the public sector). The report also identifies technology enablers of GovTech, such as internet connectivity, robust identification systems and digital signature. While technology enablers are in place, India faces teething problems with analogue enablers. The government has inspired citizens to develop digital skills through several short-term skill development courses. However, it is behind in terms of regulations. It has also taken long strides in fostering innovation in the public sector with its Startup India programme; it provides the government with feedback and helps develop solutions for the public sector space.

We can note that India’s progress is primarily on the digital governance front. The seamless integration of different platforms is yet to happen. The good news is that India is all for it. India took centre stage in the ‘Digital Governance’ or ‘GovTech’ movement, creating and successfully implementing Aadhaar, arguably the world’s largest digital identity platform. Aadhaar, along with the Digital India Mission, fuelled the development of the India Stack, which laid the foundation for integrated online public service delivery for the country. The shared digital infrastructure is essential to make governance citizen-centric. It can also help spur entrepreneurs who build solutions on this shared digital infrastructure. The Ministry of Electronics and Information Technology (MeitY) spearheads an enabling GovTech ecosystem that leverages digital platforms for transformative social, economic and governance. The changes are visible across the country. Several states have opened incubators to enhance the startup economy. They are not purely for the government as yet. However, several state governments hold startup hackathons by providing their problem statements and inviting the public to share innovative proposals. Thus, empowering other entities (such as startups and established firms) enables an open ecosystem to solve people’s governance woes. The government rewards good proposals and provides seed funds and office space for entrepreneurs to develop their solutions.

DIGIT is another platform developed by the eGovernments Foundation. The foundation was established in 2003 by Srikanth Nadhamuni and Nandan Nilekani to work with city administrators. It is an open-source and open-API powered by “developers, enterprises and citizens to build new applications and solutions. The ready-to-use platform helps achieve a quicker implementation time frame and helps local governments achieve process improvements, accountability and transparency at various levels of administration. 10 It is a manifestation of Societal Platform thinking, a systemic method to resolve complex societal challenges with speed, at scale, sustainably”. Its smart, collaborative solutions empower govern- ments and citizens to meaningfully interact with each other and catalyse urban development. 10

Hurdles to the GovTech Vision

The primary hurdle for GovTech is both technological and systemic. The gov- ernment needs to integrate data from all departments starting from the tech- nology front. It also needs to get people to allow seamless data sharing across all government departments from the systemic front. While the technological front is seemingly effortless, changing people is difficult. The GovTech approach requires open digital ecosystems and the government is moving towards the open platform vision. The UPI was an open platform that allowed several other payment systems access and launch their apps. The government also allowed third parties to access Aadhaar using APIs for eKYC purposes. Apart from technological barriers, there are several systemic barriers. 3 First, the public sector lacks technological under- standing and finds it challenging to recruit and retain technical talent as good as those in the private sector. Second, government departments are not always so eager to work with startups. A tech-savvy head of the department may do so, but not all are willing to engage with them. Third, government procurement is cumbersome and slow. Fourth, the already-established relationships and contracts with technology giants become a hurdle in engaging startups. The public sector tends to over-rely on individual digitisation champions. Finally, the public sector generally does not incentivise experimentation.

Building GovTech through AI-driven solutions

Artificial intelligence (AI) is the buzzword today, and many startups are trying to develop AI-driven solutions for governments. According to Accenture, 92 per cent of US citizens think their government has improved the use of digital services; the government can save 96.7 million - 1.2 billion hours annually, auto- mating their tasks and reducing paperwork for employees. 2 Several countries have implemented AI in GovTech. 2 For example, Surry Municipal in Canada imple- mented chatbots (virtual assistants) to help residents get answers to questions on municipal infrastructure. Using AI, the Atlanta Fire Rescue Department of the US could correctly predict 73 per cent of fire incidents in buildings. In March 2016, Australia implemented chatbots in the Taxation Office, which helped resolve 88 per cent of queries on first contact.

AI does promise unprecedented efficiency in public service delivery. Scholars and practitioners have conceptualised several AI use cases in the public sector. Some of these are as follows 11 :

The private sector is already experiencing the impact of AI and is actively building processes and customer experiences around AI. Governments have also begun exploring AI applications for public service delivery. Within the Indian context, the government needs to highlight several key indicators from the health, education and agriculture sectors as it adopts AI. By using AI, doc- tors can automate diagnostics which could also help provide health services in remote areas. In the agriculture sector, India’s per hectare cereal productivity (3,000 kg/ha) is almost half of China’s and the UK’s. Pests and diseases destroy a significant amount of produce in this sector. Precision farming using AI and IoT can help India improve productivity. It uses AI and IoT to observe, measure and respond to the condition of the crops and allows precise amounts of inputs (such as fertilisers) required for the healthy growth of a crop. Similarly, in the education sector, the student-teacher ratio is not as much as planned. The only saving grace is the internet and India’s relatively high mobile phone penetration. It has more than 1 billion mobile phone users, 600 million internet users, and 374 million smartphone users. The data rates in India ($0.24/GB) are the low- est globally, with an average speed of 6 Mbps. This implies that there is massive potential for AI technology in India.

The Tamil Nadu government, for example, has been using AI for public ser- vice delivery. It has invested in creating significant internal capacity in emerging technologies such as AI, blockchain, IoT and data analytics through its Centre of Excellence in Emerging Technologies under the Tamil Nadu e-Governance Agency (TNeGA). TNeGA and the health department have developed and launched an AI-based cataract screening mobile app. This makes cataract screening immensely scalable through social volunteers, high-school students, and good Samaritans called vision ambassadors. Other states are also launching similar initiatives with the help of the GoI. This innovation won the National AI gamechanger award from NASSCOM for the year 2022. Tamil Nadu has pioneered the use of NLP- based Tamil chatbots for improved public service delivery. Recently, it launched an AI-driven agricultural pest and disease identification system using a mobile app. The education sector also implemented an innovative AI system that uses face recognition for recording attendance.

AI Challenges from the Public Policy Perspective

Apart from the generic AI challenges, the public sector poses specific challenges to its adoption. The TAM-DEF framework (Figure 3) helps test an AI system before putting it to use in the public sector. 12 It identifies six domains of AI-related pub- lic policy challenges: fairness and equity, transparency and audit, ethics, misuse protection, accountability and legal issues, and digital divide and data deficit.

Ethics

The TAM-DEF framework categorises ethics into two sub-components: human and environmental values and privacy and data protection. These components help humans through safe AI systems.

Human and Environmental Values : Humans can deal with conflicting values; machines cannot. For example, AI must do its preferential duty towards vulner- able groups such as children, the elderly, pregnant women, the sick and others. At the same time, it must conform to human values such as kindness, dignity, compassion, respect and fairness. Similarly, if an AI system is designed to optimise the recovery of a particular mineral, it would do that without worrying about the consequent damage to the environment. Therefore, the government should design AI systems to optimise on multiple objectives.

Figure 3: The TAM-DEF Framework 12

Privacy & Data Protection: The government needs to enunciate stricter laws to protect citizens’ data. For example, the government collects location data from their CCTV (closed circuit television) feeds, eating habits, preferences regarding purchases and entertainment (movies and music), etc., that people share across AI networks. Unscrupulous people can misuse this data causing severe damage to the government.

Transparency and Audit

AI-based autonomous systems (robots) interact with people from several domains such as healthcare, finance, education, elderly care and transportation. AI systems are generally black boxes as it is difficult to understand why an AI system makes a particular decision. 13 This causes a problem from the perspective of transparency and audit. The problem becomes acute in legal cases and litigation where one needs to explain the decision for legal purposes. Therefore, the AI system should at least identify the broad contours of decision-making.

Digital Divide and Data-deficit

Having data as the foundation for the AI revolution leaves us with the danger whereby societies with poorer access to the internet, IT and digitisation will be left behind. The countries with good quality data at high granularity will benefit the most from this tsunami. And the poor communities in developing countries will suffer as they cannot invest in data. AI technology can also skew power distribu- tion between digital haves and have-nots, whereby only those with access to the online and data-driven ecosystems will be heard and have the ability to influence discourse.

Fairness and Equity

AI systems are often criticised for carrying the bias present in the training data. Historically, all societies have suffered from some form of inequity. Since AI sys- tems learn on training data, they may carry the same biases and disrupt the current social order. AI can generate new paradigms that may expose people with poorer bargaining power to another possible form of exploitation and unfair treatment. Therefore, it is crucial to design AI systems considering equity and other social values so that everyone gets equal opportunity. Fairness is another key need for such systems. As such, these systems must be trained through appropriate data to learn appropriate human values and not exhibit identity-based bias. They should also avoid social profiling, especially in crime prevention and fraud detection.

Accountability and Legal Issues

Since AI-equipped machines can take autonomous decisions from a purely unin- tended route, there arises the question of accountability. AI-equipped machines can invent different ways of accomplishing their task owing to their self-learning ability using reinforcement learning. However, this can have highly unpredict- able consequences. For example, OpenAI has created GPT2 14 – a highly accurate AI-driven text generator. However, OpenAI has released its smaller version only due to concerns about generating biased and abusive language at scale. Recently, for example, GPT4 (successor of GPT2) found a way to beat a CAPTCHA test. 15 When it recognised that it could not pass the captcha test, it hired a human to help with navigating CAPTCHA. When the human worker asked GPT4, if it was a robot, it pretended to be a person with vision impairment who could not complete the task.

Misuse Protection

Unscrupulous people may likely get their hands on AI technology and subject it to the wrong purposes. For example, although the internet benefitted billions of people across the world, criminals used it for cybercrimes, malware attacks, spreading viruses and dangerous games such as the Blue Whale Challenge that caused the loss of many lives. During the winter of 2017, FBI agents in the US monitored a hostage situation. Criminals went to the extent of using a swarm of drones to force FBI agents out of their designated spots and live-streamed the video on YouTube. 16 AI in the hands of dictatorial governments can be dangerous as they can use it to extend their unlawful regimes and suppress freedom.

These six challenges are not mutually exclusive despite being largely inde- pendent. For example, we can link the Digital Divide to Equity and Fairness. However, Digital Divide is independent of Equity as more than half of the world’s population does not have access to the internet. There are large communities still without any data to train an AI system. 17

Strategy to Overcome These Challenges

The DEEP-MAX framework identifies four steps to handle the six AI challenges identified by the TAM-DEF framework.

AI Standardisation and Rating

Since governments need to deploy AI systems worldwide, there must be a world- wide alliance for the standardisation and rating of AI systems, akin to ICANN (Internet Corporation for Assigned Names and Numbers) for the Internet. This alliance could regulate AI development across sectors. Democratic, and fair stan- dards would give confidence to users for developing and deploying AI systems. As the internet is the same standard worldwide, having a uniform standard would promote standardisation and interoperability. The alliance should define privacy standards, ethical boundaries, civil and criminal liability, and audit standards for AI systems.

DEEP-MAX Scorecard

The DEEP-MAX scorecard 12 (Figure 4) is based on the TAM-DEF framework and generates a safety and social desirability score for a given AI system using suitably designed test data sets. DEEP-MAX considers seven key parameters for point-based rating: Diversity, Equity and fairness, Ethics, Privacy and Data Protection, Misuse Protection, Audit and Transparency, Digital Divide and Data Deficit. The DEEP-MAX scores reflect the quality of an AI system. Stakeholders such as users, government departments and system integrators can use it for devel- oping, evaluating or using any AI system.

Figure 4: DEEP-MAX scorecard for AI under TAM-DEF framework ¹²

There is no one-one correlation between the TAM-DEF framework and the DEEP-MAX scorecard. For example, it does not include accountability and legal issues as these pertain to the domain of law. The issue of accountability becomes important for autonomous AI systems, where a decision may be irre- versible. We also do not include legal issues because legal systems do not lend themselves easily to mathematical treatment, hence developing a uniform score is difficult. Moreover, they differ across nations. We also split the Fairness and Equity component of the TAM-DEF framework into Diversity and Equity. Similarly, we split the Ethics component into Privacy and Ethics. As these are critical concerns, we must consider them separately. For example, it is important to train AI modules on diversity before any AI system interacts with people or makes any decision about them. Similarly, privacy is the most critical concern under the ethics component.

Blockchain for safe and TAM-DEF-compliant AI

Verifying an AI system for a safe and socially desirable AI solution is important. We can do the same using blockchain technology. An AI Certification Transparency & Scorecard Blockchain (ACTS-B) can capture the metadata about the AI-training dataset and track if the training dataset satisfies important criteria such as diversity and equity. The ACTS-B will also store the DEEP-MAX score for a given AI system. Although blockchain is still under experimentation, it can provide a secure mechanism to conduct AI-based transactions. First, and very importantly, it can provide a trusted mechanism to verify the quality of training data. Second, it can provide tamperproof DEEP-MAX scores which can be combined with certificates for training data. Third, it will carry infor- mation about AI systems’ activation atlas along with outputs of the first and second points above. The activation atlas visually presents the internal neural net nodes as features to help improve our understanding of the AI systems’ decision-making process. We can use this information from the activation atlas to alert users about the potential pitfalls of the AI model. Finally, it can help prevent misuse through its audit trail. As criminals can tweak an AI-based face recognition system for unscrupulous use, a Blockchain-driven record keeping for modifications in the criminal image database will keep the system safe from unwarranted and unsuspecting abuse.

Periodic update of DEEP-MAX Scores of AI modules in public use

Since many AI systems self-learn, the DEEP-MAX scores should be periodically updated. The periodicity of the update depends on the class of AI’s case and the degree to which it can make autonomous decisions. Most AI programs will be used as off-the-shelf components that can be assembled to develop more sophisti- cated AI. If a poorly designed AI component (e.g., for diversity) is used in real-life scenarios such as crime prevention, the results can be devastating. Such a system can lead to bias of all kinds, denying financial access based on pincode or local- ity, or even unfair charging of higher insurance premiums on account of specific characteristics.

Conclusion

India is emerging as a leader in innovating with new technologies, not just in the private sector but also in the public sector. Still, it has a long way to go. Solutions like UPI, Aadhar and Co-WIN platforms have created a new bench- mark for GovTech use cases. As the public sector embraces AI-driven solutions, it has to start worrying about potential pitfalls and issues around safety, equity and inclusion. In this chapter, we discussed a framework (TAM-DEF) and a scorecard (DEEP-MAX) to guide the governments and public agencies in ensuring that the AI systems being rolled out for public use are safe and have minimum unin- tended consequences. Regulating AI transparently is a big challenge and requires cross-national cooperation among governments. Creating test data sets for the DEEP-MAX framework and rating the AI systems on this scorecard requires a lot of research and regulatory and industry collaboration. We suggest using the DEEP-MAX scorecard in binary mode in the beginning and evolving on each of the seven parameters within the context of the proposed AI use cases, as the global standardised test datasets become available.

References

• GovTech: Putting People First, The World Bank (Last accessed on 24 March 2022)
• GovTech / E-governance Global Industry Landscape in the Post-Pandemic World 2021/Q2 (Last accessed on 24 March 2022)
• Brian Diaz (2021), Where is the Future of GovTech Heading? (Last accessed on 24 March 2022)
• Jeff Puckett (2020), GovTech reimagined in the COVID-19 era (Last accessed on 24 March 2022)
• Tanya Filer (2019), Thinking about GovTech: A Brief Guide for Policymakers (Last accessed on 24 March 2022)
• Integrated Public Governance (Last accessed on 24 March 2022)
• How governments use GovTech to fight COVID-19 (Last accessed on 05 September 2023)
• World Bank Homepage
• World Bank Group, Digital Dividends (Last accessed on 24 March 2022)
• DIGIT
• Cem Dilmegani (2023), AI in Government: Examples, Challenges & Best Practices (Last accessed on 24 March 2022)
• Dwivedi et al., Artificial Intelligence (AI): Multidisciplinary perspectives on emerging challenges, opportunities, and agenda for research, practice, and policy, International Journal of Information Management, Volume 57, 2021.
• The Economist, Advanced science and the future of government (Last accessed on 24 March 2022)
• Better language models and their implications (Last accessed on 24 March 2022)
• TN Viral Desk (2023), GPT-4 pretends to be blind person to beat CAPTCHA test, hires human to solve it (Last accessed on 24 March 2022)
• Victor Tangermann (2018), Criminals are now using swarms of small drones to befuddle law enforcement (Last accessed on 24 March 2022)
• World Bank Data Indicator: Internet Users (Last accessed on 14 April 2019)