Why are governments struggling to adopt AI?
Many governments around the world are exploring options to use Artificial Intelligence (AI) across public services. These governments see compelling reasons to use AI ranging from:
Enhanced decision-making and policy formulation. AI tools analyzing large volumes of data to identify trends, predict outcomes, and provide insights that support evidence-based policy-making. Governments can use these insights to improve the effectiveness and efficiency of public programs and policies.
Improved public service delivery. AI-driven automation that could streamline and enhance public services by reducing manual processes, minimizing errors, and enabling faster response times. For example, chatbots and virtual assistants can improve citizen engagement and provide 24/7 support for various services.
Cost reduction and operational efficiency. AI that can optimise resource allocation and reduce operational costs by automating repetitive tasks, detecting fraud, and enhance procurement processes. Predictive maintenance for public infrastructure and systems could also reduce downtime and expenses.
Data-driven governance and risk management. AI enhancing the ability to monitor and assess risks in real-time by processing vast data from various sources. Governments can use this to predict and mitigate risks, such as natural disasters, financial crises, and public health emergencies.
Fostering innovation and economic growth. AI adoption to stimulate innovation in the public and private sectors, leading to job creation, improved productivity, and overall economic growth. Governments investing in AI can also position themselves as leaders in technology-driven governance and attract investment and talent.
Given all of the possibilities, why are governments across the world struggling to adopt AI and gain benefits from its huge potential?
In a 2023 article, AI adoption and diffusion in public administration - A systematic literature review and future research agenda, Madan, and Ashok discuss the implementation of AI in public administration. The article highlights the barriers to government adoption of AI and proposes further research on AI adoption, implementation, and ideas to address identified barriers. Madan and Ashok note the four primary areas to AI adoption across public administration as being related to:
The technology context
The IT assets and capabilities required for AI adoption include cloud computing, digital infrastructure (connectivity, bandwidth, processing power), system compatibility, and data integration capabilities. Alongside this is the accessibility, quality, and management of data assets, along with cloud storage and cross-agency data sharing.
It was also found that IT capabilities focus on managing these assets, AI knowledge, and 'data-driven' culture are crucial as well. Specialized expertise is noted as frequently being a pre-requisite, often sourced through external partnerships due to the scarcity of in-house AI experts. The perceived benefits of AI, such as cost savings and innovative solutions, drive adoption.
Organisational context
The organisational context is shaped by three areas: culture, leadership, and inertia. Madan and Ashok found that innovative cultures are more receptive to AI adoption, especially in environments open to experimentation despite high risks.
Leadership also plays a crucial role, particularly through transformational leaders who drive AI initiatives by fostering innovation and motivating employees to embrace new technologies. CIOs with both technical and political expertise are hence essential for AI integration. Organisational inertia, whether due to rigid structures, centralized decision-making, or resource constraints, acts as a significant barrier to AI adoption.
Environmental context
The environmental context includes vertical and horizontal pressures. Vertical pressures involve government mandates and policy directives promoting digitalisation and AI adoption. Examples include national AI strategies and regulatory guidelines. Horizontal pressures arise from intergovernmental competition, citizen demanding better services, industry innovation, and media scrutiny. These factors create a compelling case for AI adoption in public administration, driven by the need for improved performance, cost savings, and meeting public expectations.
Absorptive capacity
Absorptive capacity cuts across all contexts and refers to an organisation's ability to integrate AI technologies. This capacity is built on existing e-government infrastructure, inter-organisational collaborations, and networks of external experts. Effective knowledge management, dynamic capabilities, and experience with deterministic systems help organisations align AI adoption with public value goals, responding to both internal needs and external pressures.
Factors impacting AI adoption across Public Administration
Technology context |
IT Assets · Cloud computing capabilities · Current digital infrastructure: high connectivity and bandwidth, processing power and server hardware, networks, system integration · Compatibility of existing assets · Data quality, availability, accessibility · Database management infrastructure · Data ownership and sharing · Storage – cloud or on-premises · Data governance maturity · Enterprise architecture
IT capabilities
· Current capabilities in managing IT assets · Staff's knowledge of AI and big data · Data-oriented culture · Big data and analytics specialists and experts · Ecosystem of commercial partners and experts
Perceived benefits
· Expected benefits · Simple intuitive design · Users' needs · Direct benefits of costs and novel solutions · Indirect benefits of increased collaboration with peers and industry |
Organisational context |
Organisational
Organisational culture
· Innovativeness, risk-taking, experimentation · Institutional arrangements such as NPM orientation, e-government · Technology and strategy alignment, cross-agency collaborations
Leadership
· Transformational leadership, institutionalising learning, and experimentation · CIO's leadership and technical expertise
Inertia
· Bureaucracy and centralised decision-making · Status-quo bias · Lack of employee empowerment · Resistance to data sharing · Resource scarcity · Cost versus benefits for experimental projects · Resistance from unions |
Environmental context |
Environmental context
· Political environment, election cycles · Policy signals, directives, mandates · Regulations, laws, procurement practices · National AI guidelines
Horizontal pressures
· Inter-governmental competitive pressures · Media scrutiny and oversight · Citizen demands · Industry pressure
Absorbative capacity
· Path-dependency · Knowledge management practices · Dynamic capabilities
|
Conclusion
Governments worldwide recognize the immense potential of AI to enhance decision-making, streamline public service delivery, reduce costs, and foster innovation. Despite this, they face significant barriers in adopting AI effectively. Key challenges span technology, organisational culture, environmental factors, and absorptive capacity.
The technological challenges include gaps in digital infrastructure, data management, and system compatibility, often compounded by the lack of in-house AI expertise. Organisational resistance, rooted in rigid structures and leadership dynamics, further complicates AI integration. Leaders play a critical role in driving AI initiatives, yet the absence of an innovative culture or transformational leadership often hinders progress. Additionally, pressures from both vertical (government policies) and horizontal (citizen demands and industry competition) directions create a complex environment for AI adoption.
The concept of absorptive capacity, which cuts across all these challenges, is crucial for integrating AI technologies. This capacity is influenced by factors such as existing e-government systems, collaboration networks, and the ability to manage and apply new knowledge. Governments that effectively align these elements can better navigate the barriers to AI adoption. However, realizing the full benefits of AI requires not just overcoming these technological and organizational hurdles, but also fostering a culture that embraces innovation while balancing ethical concerns and public value goals. As Madan and Ashok’s study suggests, addressing these multifaceted challenges will be key to advancing AI adoption in public administration and unlocking its transformative potential.
References
Madan, R., & Ashok, M. (2023). AI adoption and diffusion in public administration: A systematic literature review and future research agenda. Government Information Quarterly, 40(1), 101774.
Σχόλια