An Ambiguous Relationship between Public Administration and AI

Abstract

This chapter is an attempt to present how artificial intelligence (AI) impacts public administration and the basic ethical challenges it brings to this environment. The introduction shows the main model of relation between public administration and AI and also gives the main questions that will guide all argumentations. The second section sets up the framework for further analysis and it explains the rudimental for the whole analysis term of public administration and AI. In the second section, the fields of the use of AI in public administration are shown, including the decision-making process and other forms of public services. The third is dedicated to ethical issues that are inextricably linked with administration and AI usage.

1. Introduction

A constantly changing world affects every aspect of our lives. The most visible sign of the new time we live in is new technologies, which are increasingly present in our lives and become an unnoticeable and natural part of our existence. With the same impact, artificial intelligence (AI) impacts public administration, in which it becomes one of the tools enhancing governing an increasingly complex extra-legal reality. The general picture of this problem shows that relation between administration and new technologies is multi-faceted. On the one hand, public administration is the user of technological advances. On the other, it can be a regulator of new technologies, which plays the main role in deciding the direction of development in new technologies and the ways they can be introduced into our lives. All these lead to the basis for a multi-dimensional analysis of the subject. First of all, as the user of AI public administration can increase an impact on the lives of individuals, by making them more comfortable but also more controllable. Digitalization and automatization of administrative procedures, supporting realizations of smart city ideas, help with creating a more democratic, friendly, and effective administration. At the same time, even with the same tools, public authorities can become all-powerful decision-makers about our rights and responsibilities, which can breach the need for transparency, privacy, and discrimination. Digitalization of governing bodies can end in making administration and algorithm, which will make social participation in governing illusionary, same and change the whole model of public administration.

The above issues open a new perspective for scientific analysis of possible relations between public administration and AI. It can be described and evaluated through an explanation of the following subjects:

1. how should we understand public administration and AI?

2. in which fields AI is used in public administration?

3. what kind of rules regulate the use of AI in actions of public administration?

4. is it possible to talk about the ethical code in an administration that regulates AI?

2. The meaning of public administration and the AI

To examine the relation between public administration and AI it is essential to explain what should be understood as these two parties of this relationship. Establishing it seems to be crucial for identifying possible relations and the impact these two would have on each other.

When it comes to public administration first intuitions usually connect this term with state and authorities, administrative bodies [1]. This approach sees it as an organization, which is built with central and local administration bodies that perform the will of the states. In this perspective, administration is every government at the central level and local authorities on the territorial, and local. We can call this approach subjective or static, which emphasizes systemic matters.

Administration is also perceived as a process or activity of government, government corporations, and specialized agencies [2]. This prospect can be seen as dynamic, in which administration is an action that provides welfare and social security services to individuals. It meets the contemporary vision of public administration in which it is performing the administration is connected with democratic citizenship, public interest, public policy, and services to society [3].

The combination of those two perspectives is visible in the definition of administration constructed by Nigro and Nigro in which it is a cooperative group effort in a public setting, which has an important role in the formulation of public policy and is a part of policy, differs from private administration, and provides services to community by covering executive, legislative, and judicial, and their interrelationships [4].

Thus, public administration is the organization that has a public character and is a part of government (or separate local) power that is dedicated to pursuing public interest through executive, legislative, and judicial actions [1, 4].

Similar to public administration also AI is an ambiguous term, which is mixed aspects of its: structure, behavior, capability, function, and principle [5, 6] of intelligent machines or software.

As a system, AI is defined as software that can achieve complex goals in complex environments and can process information and adapt to its environment with sufficient knowledge and resources [6, 7]. In this approach, AI mimics the human brain and its activities [6]. It can be named behavior because through the processing it adapts to goals in a range of environments and can be said to be intelligent and act appropriately taking appropriate actions at the same time [8, 9]. Functioning like this, AI behaves like a human-like mind without demanding a human-like brain understood as a source of intelligence [6]. Capability is understood as a problem-solving skill, which is possible to achieve thanks to the cognitive functions of AI that are identified in the human mind. The principle of AI defines focus of group on fundamental principles of the usage of intelligence as a form of rationality. In this matter, the main issue is an explanation AI can make the best possible decision in various situations [6]. Therefore, it can be named as an intelligent system with perception, process, decision-making, learning, and adaption abilities supported by dedicated software or hardware [10] or “a cluster of digital technologies that enable machines to learn and solve cognitive problems autonomously without human intervention” [11].

Comparing these two actors in the relation it is visible how many common points they share. Both AI and public administration can be considered as systems that provide us with decisions in uncertain and often unique circumstances. This conclusion indicates the main connection between them, which makes AI useful for administration and its actions.

3. The public administration and AI: the main correlations

Although the adoption of AI solutions in various business sectors is dynamically emerging, it is still in its early stages in public administration, where it is still considered with suspicion. Therefore, it is hard to say it is a true game changer in the public sector. Especially, that with it comes not only technological change, but also transformation and innovation of culture, processes, and workforce [12]. Still, it is constantly changing with the approach determined by nowadays world where the most important are efficiency and cost savings, which makes the stage for AI solutions that provide both. This makes AI and its usage in the public sector not only novelty and fashionable but also necessary innovation to make the administration keep up with times and external reality [13]. It is called a “response to the growing complexity of the public sector, and a need to handle larger and more advanced forms of data” [14]. The diversity of AI software allows it to be applied in every action of public administration. Among others, AI helps to keep up with the standards and principles that characterize contemporary society.

Therefore, AI becomes more present in internal and external relations in which public administration is involved. With this, it changes both the system and the ways of performing administration, and the lives of citizens.

The application of AI in public administration can be done in each aspect of its activity, in (1) policymaking, (2) public service delivery, and (3) internal management of public organizations [15] in every field of public services that are covered by public administration: including healthcare, transportation, and education [16]. To achieve effects it is possible to use: (1) automated decision-making applications, (2) virtual agents or (AI) chatbots, and (3) applications related to public governance and public security [17].

The use of AI is especially significant in decision-making processes conducted in administration. It can play different roles. First, it can be the tool to gather the material needed to make a final judgment or it can be the tool of exact decision-making.

The biggest advantage of AI in public administration is its ability to enhance operational efficiency, even when it is not responsible for decision-making. AI systems can process vast amounts of data at unprecedented speeds, enabling quicker decision-making and reducing the burden on human employees [18]. AI includes process automation, virtual agents and speech analytics, predictive analytics for decision-making, sentiment analysis, and document reviews [19].

For instance, AI-driven algorithms can analyze demographic data to predict and manage public service demands, such as healthcare or transportation needs, more effectively [20]. Moreover, AI can support decision-making processes by providing data-driven insights. Machine learning models can identify patterns and trends that may not be immediately apparent to human analysts [21]. This capability is particularly valuable in areas such as urban planning, where AI can help predict population growth, traffic patterns, and environmental impacts, thereby informing more strategic and sustainable development plans. Same as in car traffic.

What is the most significant is the automatization of the process of application of the law, in which slowly AI can replace the human factor. AI can enhance the efficiency of administrative processes. For example, automated systems can streamline tasks such as processing applications for permits or licenses, reducing wait times, and minimizing errors. In the healthcare sector, AI can assist in diagnosing diseases, managing patient records, and even predicting outbreaks of epidemics, thereby improving the overall quality and accessibility of public health services [17].

In the light of earlier conclusions, due to the technological revolution we can now it is possible to distinguish three forms of decision-making in public administration: (1) no decision automation, where humans keep full the decision authority without the use of algorithms; (2) semi-automated decisions, in which human is still the primary decision-makers, what he does aided by algorithms that present data, propose procedural actions, or offer a set of potential decisions; and (3) fully automated decisions, where algorithms hold the decision-making authority, operating without humans. AI-based decision-making is perceived as a techno-rational free from human impact, which is at the same time objective and neutral [22]. What is more, it increases the transparency of decisions and reduces the risk of corruption in public administration [23], because it gathers, analyzes, and leverages data to make automated decisions, with minimal human involvement [24].

One of the first examples of an automated public decision-making process is from Canada, which in 2014 began using automated decision-making in sorting and filtering. The reason to do so was the increasing number of applications that were to be engaging for “manual” operations [22]. AI makes all the processes faster and both impartial and fair. It leads to the conclusion, that the rudimental for such a proceeding was to ensure the ethical and fair use of AI. To do so, the Government of Canada published a White Paper on Responsible Artificial Intelligence on April 1, 2019, which ensures the transparency, accountability, legitimacy, and fairness of automated decision-making, and imposes certain requirements on the use of decision-making algorithms and systems by the government.

The White Paper is the first such act in history and has been in force since April 2020. The foundation for this regulation is algorithmic impact assessment which is the responsibility of the agencies before an automated decision-making system is set up or changed. To ensure proper use of AI in making decision processes, it includes questions such as whether customers belong to a vulnerable group, how complex the algorithm is, who collected the data used to create the system, whether the system replaces a human decision, and if necessary whether the system can justify its decisions, whether man can override the decisions, and whether the effects of the decision can be reversed. Managers are required to publish the results of their assessments online as open government data making them transparent and achievable for everyone. Thanks to these conditions automated decision programs are intended to be comprehensive, publicly available overviews of existing [16]. What is more, automatization makes the whole process more predictable.

Predictability or stability is needed in the public sector, but decision-making models are not only based on algorithms, especially when it comes to discretion. It is undeniable that when the procedure is highly formalized and its effect is determined only totally with the content of the norm, as it is in declaratory decisions, determined only by clear and unconditional circumstances indicated in the regulation, AI can be helpful. Making decisions based not only the strict norms but also on the effects of interpretation of vague concepts or balancing values is impossible to do with machines and software. For that, we need humans with experience, sensitivity, and ethics. What is more, the decision-making process depends on the environment, which is in constant movement. Internal changes such as the implementation of new policies or external ones such as new legislation, a changed environment, epidemics, and so on impact public administration and should also impact the AI decision-making processes. They can impact the trajectory of the procedure and are responsible for making circumstances individual and unique. Therefore, artificial intelligence and machine learning are most valuable when used to support, and not replace human decision-making [22].

Besides the role of AI in administrative procedures, it can also be extremely helpful in generally understood public services. It undoubtedly improves the delivery of public services. Chatbots and virtual assistants, powered by natural language processing, are increasingly being employed to manage routine inquiries and provide information to citizens. These AI-driven tools can operate 24/7, offering timely assistance and freeing up human resources to focus on more complex tasks [25]. Their friendly interfaces and intuitive use make them reachable and easy to use. In the already cited Canadian White Paper [22, 26], AI brings significant gains in public services. As it is indicated: (1) smarter search: Natural language processing can advance the public sector’s user interaction and then profile the services when the algorithms will learn patterns to better understand what citizens want when accessing government’s services; (2) the use of chatbots, which can help filter routine questions away from human service agents so that they may focus on helping users through more complex or cases, which can rely on machine and software; and (3) automated decision support increasing quality of service by cutting wait times.

Still, they are not humans, which is why such experiments often meet mental barriers. However, some groups see AI as the best tool, which is more trustworthy than human service, because it is indifferent to race or gender which may lead to discrimination [27]. Still at the same time large group of respondents would have problems with trust when human services were replaced by chatbots [28].

It cannot be denied that the success of AI in public administration depends on citizens’ perceptions of AI in public decision-making and other actions of administrative bodies. AI influences citizens’ lives and whole societies, therefore so important are factors of trust and legitimization of its action. This is provided by giving society and individuals AI tools to take part in administration and monitoring. This gives AI the potential to enhance social participation, which refers to the involvement of individuals in community activities and decision-making processes, which is crucial for a healthy and democratic society. AI can play a significant role in increasing social participation by improving accessibility, fostering inclusivity, enhancing civic engagement, and facilitating better communication and coordination.

Civic engagement, which involves activities like voting, volunteering, and participating in public discussions, is another area where AI can make a significant impact. AI tools can streamline the voting process by offering personalized reminders and information about elections, thereby increasing voter turnout. Furthermore, AI-powered platforms can facilitate public consultations and deliberations, enabling citizens to voice their opinions on various issues more conveniently. For example, AI chatbots can be deployed to collect feedback from residents about community projects, ensuring that decision-makers receive input from a broader segment of the population.

Furthermore, AI can enhance these aspects by providing tools that enable more efficient information sharing and collaboration. For instance, AI-powered project management software can help community organizations coordinate events and activities more effectively, ensuring that volunteers and participants are well-informed and engaged. Additionally, AI can be used to analyze social networks and identify key influencers, helping organizations target their outreach efforts more strategically.

Additionally, AI fosters inclusivity by identifying and mitigating biases that often exclude marginalized communities from social participation. Machine learning algorithms can analyze large datasets to uncover patterns of discrimination or underrepresentation, allowing organizations to address these issues more effectively. AI can support the development of more inclusive policies by providing data-driven insights into the needs and preferences of diverse populations. What is more, it is improving accessibility. AI-powered tools such as speech recognition software, screen readers, and automated translation services can break down barriers for individuals with disabilities and non-native speakers. For example, AI-driven voice assistants can help visually impaired individuals navigate digital spaces, while real-time translation apps enable seamless communication across different languages. By making information and services more accessible, AI ensures that a broader range of people can engage in social, cultural, and political activities. Such solutions are necessary in the globalized world and can help with dealing with contemporary crises like migration or war in Ukraine, which made people change their environment and move to other countries. Thanks to such tools they can start and lead an existence in new surroundings, which reduces the typical pressure connected with such changes.

Implementation of AI in public administration also holds the potential to promote transparency and accountability of administration. AI systems can monitor government activities and detect anomalies or signs of corruption more effectively than traditional methods. AI can identify irregularities and flag suspicious activities for further investigation. Furthermore, AI can enhance public engagement by making government data more accessible and understandable to citizens. Such methods are crowdfunding which is possible to conduct with AI and can be a huge input to policy development and decision-making.

Open data initiatives, supported by AI analytics, can empower citizens to scrutinize government performance and contribute to policy discussions, fostering a more inclusive and participatory governance model. Which, thanks to new technology administration become more open and approachable to citizens, changes the nature of the relationship between those two into more affiliate form.

4. Ethical issues

Despite its numerous benefits as cutting the cost of administration, transparency, and efficiency, the use of AI in public administration is not without challenges. The rapid and wide adoption of AI in governmental operations and decision-making processes makes very visible the intersection of ethical challenges. The use of AI in public administration is directly connected with ethical tensions of fairness, transparency, privacy, and human rights [22]. In this perspective, ethics is not only a matter of public administration but also of the AI. It is possible to indicate the ethics of AI and the ethics of using AI in the actions of public administration.

How it sounds from earlier considerations AI can support ethics in public administration. It provides impartiality and transparency. Simultaneously itself and its use in public administration can be the biggest threat to ethics. Therefore, it is important to create ethical AI and ethical rules determining the use of it in public services. It starts with the design of algorithms to the implementation of them into policies and adopting regulations.

The significant concern is the issue of privacy, which is a right essential to human dignity but is also the most affected by the negative impact of AI [29]. The development of AI systems relies heavily on a huge amount of data, including personal data and private data [30]. Explaining this by acting for public safety thanks to AI solutions, public administration can synthesize data and estimate personal characteristics, such as gender, age, marital status, and occupation, from cell phone location data [29], thanks to scoring and public monitoring it helps to create facial databases [29]. Therefore, it is crucial to provide data privacy and security, to avoid misuse and malicious use [31] and the main role it should play in public administration.

From this emerges an obligation for governments to ensure that the data collected and processed by AI systems is protected against breaches and misuse. Additionally, there is the risk of bias in AI algorithms, which can lead to unfair or discriminatory outcomes. Public administrations must implement measures that ensure AI systems are transparent, accountable, and free from bias. They advocated for the development of policies to balance the benefits of AI against its potential risks, ensuring that advances in technology do not compromise democratic values and human rights.

The majority of AI literature views the government as a regulator and at the same time a significant user of AI [22]. It makes it the best source for ethical rules that can combine practice and expectation levels. As AI continues to evolve and integrate into various aspects of society, it brings forth a host of ethical considerations that necessitate robust legislative frameworks, which can be provided by public administration. To address ethical concerns connected with AI, governments and international bodies are developing and implementing laws and regulations aimed at guiding the ethical deployment of AI technologies. Through establishing clear guidelines and principles, these legislative frameworks aim to protect individual rights, promote fairness, and foster trust in AI systems. As AI continues to evolve, ongoing dialog and collaboration among stakeholders will be vital to creating a sustainable and ethical AI ecosystem.

To date, many acts with ethical guidelines for AI use have been published on international and state levels. Among others, they are for example EU’s ethical guidelines (European Commission, 2019), Canada’s Algorithmic Impact Assessment (Canada, 2020), UK’s guidance (Gov. uk, 2019b). They all provide overall boundaries for the use of AI [31]; however, international rules that would be the standard ethical code for the use of AI are still not already created. As a prosthetic act of this nature, we can treat the law enforced by the European Union (EU) or Organization for Economic Cooperation and Development (OECD).

The European Union (EU) has been at the forefront of regulating AI ethics, primarily through the General Data Protection Regulation (GDPR) and the proposed AI Act. The GDPR, enacted in 2018, sets stringent guidelines on data protection and privacy, directly impacting AI systems that process personal data. Key principles of the GDPR include data minimization, purpose limitation, and the right to be informed about automated decision-making. These principles ensure that AI systems are designed with privacy in mind, protecting individuals’ data from misuse and unauthorized access. In addition to the GDPR, the EU has proposed the AI Act, which seeks to establish a comprehensive legal framework for AI. The AI Act categorizes AI applications into different risk levels (unacceptable, high, and minimal risk) and imposes corresponding regulatory requirements. High-risk AI systems, such as those used in healthcare, transportation, and employment, must meet stringent criteria related to transparency, accountability, and human oversight. By setting clear standards, the AI Act aims to mitigate the risks associated with AI while promoting innovation and trust.

Beyond national regulations, international organizations are also playing a crucial role in shaping AI ethics. UNESCO’s Recommendation on the Ethics of Artificial Intelligence, adopted in 2021, provides a global framework for ethical AI. It emphasizes principles such as human rights, fairness, transparency, and accountability, urging member states to incorporate these principles into their national legislation and policies. By promoting international cooperation and harmonization, UNESCO aims to ensure that AI development respects human dignity and ethical standards worldwide.

The Organization for Economic Cooperation and Development (OECD) has also developed principles for AI ethics, which have been endorsed by its member countries. The OECD Principles on AI focus on ensuring AI systems are robust, safe, fair, and transparent. They call for inclusive growth, sustainable development, and well-being while emphasizing the importance of human-centered values and fairness in AI design and implementation.

The other ethical issue is the impact of AI on the administration itself. As elsewhere, it is possible in its organization potential displacement of jobs due to automatization. While AI can take over repetitive and mundane tasks as indicated before, it is essential to manage the transition for affected employees through reskilling and upskilling programs. Public administrations must strike a balance between leveraging AI for efficiency and ensuring a fair and equitable workforce transition. It is undeniable that this is a sign of reducing costs of public administration, but it also has a great impact on the citizen’s lives and the existence level of people working in administration. The displacement of administrative job positions can lead to economic challenges for affected workers, including loss of income and job security. This can have broader economic implications, such as increased unemployment rates and reduced consumer spending.

Ethical considerations should guide the implementation of new technologies. Organizations should adopt responsible automation practices, ensuring that the deployment of technology does not disproportionately impact vulnerable workers. Transparent communication about technological changes and their implications can also help build trust and support among employees.

5. Conclusions

The utilization of artificial intelligence in public administration represents a transformative shift in how governments operate and serve their citizens. By enhancing efficiency, improving service delivery, and promoting transparency, AI has the potential to significantly improve the public sector on performance and organizational levels. However, it is essential to address the associated challenges and ethical considerations to ensure that AI is implemented responsibly and equitably. As AI technology continues to evolve, its integration into public administration will undoubtedly play a pivotal role in shaping the future of governance and societies. Even now it is hard to differentiate its impact on administration from the impact on citizen’s life.

Looking ahead, the role of AI in public administration is set to expand even further. Advances in AI technology, such as deep learning and predictive analytics, will enable more sophisticated applications in various public sectors. Governments will increasingly rely on AI to address complex societal challenges, from climate change to public safety. Therefore, it is important to perceive this matter from the global or international perspective. International legal regulations on AI, collaboration, and knowledge sharing will be critical in harnessing AI’s full potential. Governments can learn from each other’s experiences and best practices, fostering a global ecosystem of innovation in public administration [32].