Preventing the Failure of Generative AI Utilization in the Public Sector: What Should the Government Do?

  • 27/02/2025
  • Category: Commentaries
No Comments
AI Utilization in the Public Sector

Author: Hosea Immanuel Latumahina

Editor: Bangkit Adhi Wiguna

AI utilization will contribute 12% of the increasing national economic growth (GDP) or around USD 366 Billion in 2030. That’s what the Indonesian government has projected and believed[1]. The Indonesian government has increasingly recognized the importance of Artificial Intelligence (AI) in transforming the public sector. In recent years, there has been a growing interest in leveraging generative AI to improve public service delivery, optimize data management, and enhance decision-making processes. This shift toward AI adoption is reflected in several national policies, including the National of Artificial Intelligence Strategy Indonesia (Stratnas) 2020-2045, which outlines AI as a core element for digital transformation.

Stratnas 2020-2045 published as a guideline for Indonesia to address digital technology challenges, including the development and utilization of AI[2]. In Stratnas, AI is included as a key priority to support digital transformation, particularly in public service. This national AI policy plan focuses on digital infrastructure development, human resource capacity building, and strengthening technology-based governance systems.

Despite the direction and targets set by Stratnas, the implementation of generative AI technologies in the public sector has not yielded concrete legal regulation and has not achieved these expectations. The pace of AI integration has been slow, and many of the essential components for AI development—such as a comprehensive digital infrastructure, regulatory frameworks, and data governance policies—are still in their nascent stages.

The gap between the strategic vision outlined in Stratnas and the actual progress made on the ground is evident in the lack of concrete AI applications in government services. Various government initiatives are still experimenting with AI, rather than fully utilizing it to improve public sector efficiency or deliver more personalized services. Furthermore, concerns over data privacy, cybersecurity, and the uneven distribution of digital infrastructure across Indonesia have made it difficult to achieve the ambitious goals set in the strategy.

Digital Public Infrastructure as a Basis of Generative AI Utilization in Public Sector

In the public sector, Digital Public Infrastructure (DPI) refers to the foundational digital systems and services that enable the effective delivery of public services through technology, including AI (OECD, 2024). DPI encompasses various components, such as digital identity systems, secure data exchange platforms, and interoperable public service applications. Without a robust DPI, the implementation of AI, particularly generative AI, in the public sector would be constrained, leading to inefficiencies and potential inequities in service delivery[3].

DPI is the shared digital systems and services that are secure, and interoperable, based on inclusive access. The three main components that support DPI are software and hardware, shared infrastructure, and services.2 These components lead DPI to be designed according to “common good” principles, which to be built as shared digital systems that are secure, interoperable, based on open standards, promote access to services for everyone, and oriented around policy priorities[4]. DPI is not just about technology, but also involves thoughtful design and governance to ensure it serves the public interest effectively.

Data management within DPI is very important. But unfortunately, Indonesia’s data management and infrastructure has not been optimally managed, often being overlapping, not integrated, and not connected. The development of digital public infrastructure must pay attention to public values and be accessible to the public, as citizens entrust their data rights to these institutions[5]. Therefore, data privacy management needs to be improved to ensure that public personal data can be accessed openly while strengthening data security measures.

Why is Digital Public Infrastructure Important?

The implementation of generative AI requires a strong DPI foundation to analyze data, train algorithms, and provide AI-based services equitably and efficiently. To make sure that the utilization is implemented well, generative-AI in the public sector must fulfill each component of Digital Public Infrastructure. Several countries have demonstrated how DPI supports the effective adoption of technology in public service. Estonia and Singapore have successfully built highly efficient DPI, supporting AI deployment in the public sector.Estonia has an excellent e-government system, with all public services accessible digitally and also leverages its integrated digital infrastructure to provide state services digitally and efficiently, including e-residency, which allows foreign entrepreneurs to access public digital services directly[6].

Estonia has built a successful digital public service through collaboration and strategic intent of participation from every actor, such as the governance, politicians, academics, business leaders, civil servants, and also citizens[7]. Furthermore, the successful case of Estonian digital development also came from the right diagnosed for its domestic challenges and problems. For example, many strategic policy documents for digital transformation from many countries have followed the rhythms of European (structural) funding periods rather than responding to domestic challenges and planning processes. Furthermore, Kattel & Mergel (2019) mentioned, as Estonia seemed to lack the political will and economic capacity to build its own significant industry, the focus turned to develop what might be labelled as a general purpose technology. In other words, rather than developing specific IT and electronics industries, the focus was on developing IT as a general purpose socio-economic skill to be shared by as many citizens as possible. This resulted in Estonian’s governance to conduct the digital transformation strategy as a leapfrog rather than just catch up with the West[8].

This digital transformation of Estonia’s governance also focuses on developing the fundamental principles such as futuristic and universal public digital architecture, and also decentralized and distributed digital agendas (including database infrastructure) of across-line government’s ministries, departments, and agencies. The government was also encouraged to develop their own digital agendas and find open-source solutions on their own rather than buying from vendors. This principle had succeeded to become an explicit strategy which enforced government’s ministries, departments, and agencies to build their IT systems according to their specific needs but ensuring frugality and interoperability across government. The resulting distributed architecture created the need for a software layer that allowed these distributed IT systems and databases to exchange data with each other[9]. However, the core of Estonia’s digital transformation that defined as the idea of a digital citizen who access public and private sectors through a digital platform had succeeded ensure the governance to increase interoperability of diverse and decentralized information systems and integrate them into crucial public-private services that concerned directly to citizen aspects (Kattel & Mergel, 2019).

Singapore, on the other hand, uses AI to optimize public services like smart city planning and data-driven healthcare systems[10]. Starting with concentrating to develop advanced data analytics and IoT devices, Singapore has succeeded in optimizing their urban planning and resource management, such as energy consumption, water usage, and waste management. Besides that, Singapore’s goverment also focusing to utilize the AI into their basic needs, such as healthcare sector which lead them to integrated telemedicine,digital health records, and AI-powered into its healthcare system, improving their accessibility and efficiency[11].

Moreover, Chkuaseli (2024) mentioned, a core aspect of Singapore’s strategy is data-driven governance, using vast amounts of data from various sources to inform policies and improve public services. The Smart Nation Sensor Platform monitors and manages public safety and environmental conditions, enabling proactive responses to issues such as public health crises and urban congestion. Also, public-private partnerships drive innovation, with collaborations between the government, private sector, and academia facilitating the rapid development and deployment of new technologies. The government strategy to provide 1000 e-services platforms like Singpass, a secure digital identity system, that included accessible services such as tax filing, passport applications, and public housing transactions, has also demonstrated digital transformation based on a DPI concept and a citizen-centric approach. These initiatives, however, have succeeded to influence an improvement on sustainable development and quality of life in this country[12].

Government Has to Reprioritized its Generative-AI’s Utilization Agenda

The inadequacy of DPI in Indonesia will significantly impact the implementation of generative AI in the public sector. Without proper infrastructure, AI deployment will be prone to algorithmic bias, which could exacerbate inequities in public service delivery. If AI models are trained on incomplete or non-representative data, the results could be misleading and harmful, particularly to underserved groups[13]. Furthermore, poor governance in the management of data and technology will worsen issues of transparency and accountability in AI usage in the public sector. This could exacerbate social disparities and erode public trust in the government[14]. For real example, a ransomware attack on Pusat Data Nasional (PDN), which issued has locked data from at least 282 ministries and local government agencies, has slowing down the process of online public services, such as arrival and departure processes, scholarship disburments, registration of foreigners for taxpayer identification numbers (NPWPs), and other disruptions[15]. Moreover, this ransomware attack on PDN’s case has decreased the public trustworthiness perception to the Government[16].

To ensure the successful implementation of generative AI in Indonesia’s public sector, it is crucial to prioritize the development of a robust. The government should invest in creating secure, interoperable systems that enable efficient data exchange and integration across various public sectors. Additionally, clear and transparent regulations on data privacy, cybersecurity, and ethical AI use must be established to foster trust among citizens and encourage wider adoption of AI technologies. Strengthening digital literacy and capacity building for public sector workers will also be essential in adapting to the AI-driven transformation.

Moreover, the government must engage in collaborative efforts with private sector players, academic institutions, and civil society to ensure that the development of AI technologies aligns with public needs and values. Lastly, a careful approach to monitoring and evaluation should be implemented to track progress and address potential risks, ensuring that the goals of Stratnas 2020-2045 are achieved within the expected timeline.

  1. Kementerian Ekonomi Indonesia, 2025. Pemerintah Terus Berikan Dukungan Terkait Optimalisasi Transformasi Digital. Kementerian Ekonomi Indonesia. Available at: https://www.ekon.go.id/publikasi/detail/5927/pemerintah-terus-berikan-dukungan-terkait-optimalisasi-transformasi-digital (Diakses: 27 Februari 2025). ↑
  2. Badan Pengkajian dan Penerapan Teknologi. (2020). Strategi Nasional Kecerdasan Artificial Indonesia 2020-2045. Badan Pengkajian dan Penerapan Teknologi. ↑
  3. Kuziemski, M., X, Y., Z. (2024). AI Governance in the Public Sector: Three Tales from the Frontiers of Automated Decision-Making in Democratic Settings. Available at: https://www.researchgate.net/publication/340725541 ↑
  4. Eaves, D., Mazzucato, M. and Pagliarini, G., 2024. Leveraging digital public infrastructures for the common good to promote inclusive and sustainable economic development in Brazil. UCL Institute for Innovation and Public Purpose, Working Paper Series (IIPP WP 2024-19). Available at: https://www.ucl.ac.uk/bartlett/public-purpose/wp2024-19 ↑
  5. Fisipol UGM (2024). Digital Society Week: Urgensi Pembangunan Digital Public Infrastructure (DPI) untuk Masyarakat. Available at: https://fisipol.ugm.ac.id/digital-society-week-urgensi-pembangunan-digital-public-infrastructure-dpi-untuk-masyarakat/ ↑
  6. Lars, E. (2024, December 4). e-Estonia – We have built a digital society & we can show you how. e-Estonia. https://e-estonia.com/ ↑
  7. Kattel, R. and Mergel, I., 2019. Estonia’s digital transformation: Mission mystique and the hiding hand. Available at: https://doi.org/10.2139/ssrn.3443374 ↑
  8. Burlamaqui, L. and Kattel, R., 2016. ‘Development as leapfrogging, not convergence, not catch-up: Towards Schumpeterian theories of finance and development’. Review of Political Economy, 28(2), pp.270–288 ↑
  9. Laar, M., 2010. Interview with Stephen J. Dubner, for ‘Freakonomics Radio’ podcast, 24 March ↑
  10. Katal, N., 2024. AI-driven healthcare services and infrastructure in smart cities. In: Smart Cities. CRC Press, pp. 150-170. ↑
  11. Chkuaseli, 2024. Singapore: A leader in smart city initiatives and digital government services. Available at: https://eustochos.com/singapore-a-leader-in-smart-city-initiatives-and-digital-government-services/ ↑
  12. Chan, J.J. and Chye, S.W.C., 2023. Impact of smart city initiatives on urban planning strategies in Singapore: An in-depth analysis of technology-driven solutions and their influence on sustainable development and quality of life. Journal of Strategic Management, 7(7), pp.11-21. ↑
  13. Anderson, 2020. The inadequacy of DPI in Indonesia and its impact on generative AI deployment in the public sector. Available at: https://surface.syr.edu/cgi/viewcontent.cgi?article=2831&context=honors_capstone ↑
  14. Corwin, S. and Lewis, L., 2019. Poor governance in the management of data and technology exacerbates issues of transparency and accountability in AI usage. Brookings Institution. Available at: https://www.brookings.edu/articles/algorithmic-bias-detection-and-mitigation-best-practices-and-policies-to-reduce-consumer-harms/ ↑
  15. Rangkuti, 2024. Dampak yang ditimbulkan dari peretasan data nasional oleh ransomware 3.0 tahun 2024. Available at: https://fahum.umsu.ac.id/blog/dampak-yang-ditimbulkan-dari-peretasan-data-nasional-oleh-ransomware-3-0-tahun-2024/ ↑
  16. Wardani, V.G., Santosa, H.P. and Setyabudi, D., 2022. Pengaruh terpaan berita kebocoran data penduduk dan terpaan negative e-word of mouth di media sosial terhadap tingkat kepercayaan masyarakat pada pemerintah pusat dalam menangani kasus kebocoran data. Interaksi Online, 11(1), pp.326-336. ↑