Public administration is the art and science of managing public programs and policies and coordinating bureaucratic strategies and public affairs. Public administration was an important part of my PhD in Political Science, and I was particularly interested in the role of information technology (IT) and networking, including its use in financial tasks. As we move forward with IT and even AI in government, it is critical that they be designed and programmed with ideals gleaned from years of public administration deliberations.

The debate over whether government should be run like a business has been a long-standing issue in public administration. The historical writings of public administration offer compelling reasons why government is fundamentally different from business. Scholars such as Paul Appleby, Harland Cleveland, Dwight Waldo, Max Weber, and even US President Woodrow Wilson have articulated key differences between government and business, emphasizing the distinct purposes, structures, and constraints that define the public administration of government. Also important is the role of politics, a fundamental component of the democratic agenda, but one that is not always conducive to efficiencies and values present in the private sector.

This post covers some of the primary reasons why government differs from business, drawing on historical writings in public administration, including political constraints, public interest vs. profit maximization, accountability and transparency, decision-making and efficiency constraints, monopoly vs. market competition, legal and ethical constraints, and distinctions between service vs. the consumer model.

By carefully considering these challenges and drawing on the wisdom of the classics of public administration, it may be possible to start to train the power of AI to create “smart” but ethical government systems that serve the public interest and promote the well-being of all citizens. Currently, the Trump administration with Elon Musk seems to be building a “digital twin” of the payment system at the Treasury and other parts of the administrative heart of the US government, probably in the new datacenter called Colussus built in Memphis.

Digital twins are a powerful tool for training AI models as they can help to generate data, simulate scenarios, and explain AI models. It mimics current systems as it trains a new AI engine with the goal of developing new types of digital bureaucracies and services. As digital twin technology develops with faster chips and larger data centers, it will likely play an even greater role in training AI government models. This innovation is new and unprecedented and should only be pursued with the highest intentions and a solid basis in democratic and public administration understanding.

Political Constraints and Efficient Bureaucracies

Woodrow Wilson (1887), in “The Study of Administration,” addressed the issue of government efficiency and argued that public administration should be distinct from politics. For him, government is ultimately driven by the public good, not financial gain. He emphasized the need for a professional and efficient bureaucracy to implement public policy. Wilson’s emphasis on the separation of politics and administration highlighted the need for a professional and impartial bureaucracy.

Paul Appleby (1945) reinforced this position by stating that government serves a broad public interest rather than a select group of stakeholders. Government’s core purpose is to serve the public interest and promote the general welfare of society. This includes providing essential services, protecting citizens, and promoting social equity.  

Governments often operate with a longer-term perspective, considering the needs of future generations and the long-term sustainability of policies and programs. Businesses, while also concerned with long-term success, often prioritize shorter-term financial goals. Businesses prioritize profit, efficiency, and shareholder value, whereas governments must balance equity, justice, and service delivery even when it’s not profitable (e.g., social security, public education). For example: The government provides social services like healthcare for seniors, unemployment reflief, and welfare, which businesses would find unprofitable.

Businesses are legally required to maximize profits for their shareholders. In contrast, government’s core purpose is to serve the public interest and promote the general welfare of society. This includes providing essential services, protecting citizens, and promoting social equity. By keeping Appleby’s insight at the forefront, AI development in government can be guided by a commitment to serving the broad public interest and strengthening democratic values.

Accountability, Transparency, and Legitimacy

Max Weber emphasized that government agencies operate under legal-rational authority, meaning they follow laws, regulations, and procedures that are meant to ensure transparency and accountability. Businesses operate under market competition and corporate governance, where decisions can be made with greater discretion without public oversight. Weber’s work on bureaucracy underscores the importance of formal rules, clear procedures, and hierarchical structures in government organizations. This translates to AI systems needing well-defined architectures, clear lines of authority for decision-making, and specific functions for each component. These frameworks may ensure accountability and prevent AI from overstepping its intended role.

In his seminal work, Economy and Society (1922), Weber articulated fundamental differences between government and business.
His analysis highlighted the structural, operational, and accountability-based distinctions between the two domains. He distinguished government from business in several ways: Government bureaucracy operate under legal authority, meaning it follows a fixed set of laws and regulations. Business bureaucracy is primarily driven by profit motives and market competition, with more flexibility in decision-making. Government officials also follow formal rules and legal mandates, while business executives can make discretionary decisions based on market conditions. For example: A government agency must adhere to strict procurement laws when purchasing supplies, whereas a business can choose vendors based on cost efficiency alone.

Dwight Waldo (1948) in The Administrative State highlighted that government accountability is complex because it must answer to multiple stakeholders (citizens, courts, legislatures), unlike businesses that primarily answer to investors. For example, governments hold public hearings and legislative reviews before making budgetary decisions, whereas businesses do not require public approval before adjusting financial strategies.

Waldo challenged the traditional view that public administration could be purely technical and neutral. Governments are accountable to the public and operate under greater transparency requirements than businesses. This includes open records laws, public hearings, and legislative oversight. Public officials are also held to higher ethical standards than private sector employees, with expectations of impartiality, fairness, and integrity in their decision-making.  

Waldo argued that bureaucracy is not just an administrative tool but a political institution, shaped by values, ideologies, and democratic principles. This makes accountability more complex than in business, where efficiency and profit are the primary concerns. His main points were:

– Bureaucracy is inherently political, not just administrative.
– Government agencies must answer to multiple, often conflicting, stakeholders.
– Bureaucratic power must be controlled through democratic institutions.
– Efficiency must be balanced with justice, ethics, and public values.

Governments possess coercive power, including the ability to tax, regulate, and enforce laws. Businesses, while also subject to regulations, primarily rely on market forces and voluntary transactions. Governments derive their legitimacy from democratic processes and the consent of the governed. Businesses, while also subject to societal expectations, primarily focus on satisfying customer demand and generating profits for investors.

Decision-Making and Efficiency Constraints

Herbert Simon (1947) in Administrative Behavior introduced the concept of “bounded rationality,” challenging the notion of perfect rationality in decision-making and explaining that government decisions are constrained by political pressures, competing interests, and complex regulatory environments.

Bounded rationality is often considered a more realistic model of human decision-making in organizations, recognizing the inherent limitations individuals face. Understanding bounded rationality can inform organizational design, promoting structures and processes that support effective decision-making within these constraints. Developing decision support tools and technologies can help overcome some of the limitations of bounded rationality, providing decision-makers with better information and analysis.

This concept recognizes that individuals, particularly in organizational settings, face inherent limitations preventing them from making perfectly rational decisions. These include limitations due to limited cognitive capacity and the inability to process all available information or consider every possible alternative when making decisions. Decision-makers also lack complete information about the situations, the potential consequences of their choices, or the preferences of others involved. Individuals are also prone to cognitive biases, such as confirmation bias (seeking information that confirms existing beliefs) and anchoring bias (over-relying on the first piece of information received), which can distort their judgment.

Simon argued that officials often “satisfice” instead of optimize. They make “good enough decisions” due to these limitations. They often choose the first option that meets their minimum criteria, rather than searching for the absolute best solution. Satisficing is often a more efficient approach, as it conserves cognitive resources and allows for quicker decision-making. However, it may not always lead to the optimal outcome.

By acknowledging the limitations of human rationality and designing AI systems that work within those constraints, governments can leverage AI to make more informed, efficient, and effective decisions. It’s about creating AI that assists human decision-makers in navigating the complexities of the real world, rather than attempting to achieve an unrealistic ideal of perfect rationality.

By acknowledging the limitations of human rationality and designing AI systems that work within those constraints, governments can leverage AI to make more informed, efficient, and effective decisions. It’s about creating AI mechanisms that assist human decision-makers in navigating the complexities of the “real world,” rather than attempting to achieve an unrealistic ideal of perfect rationality.

Philip Selznick (1949) in TVA and the Grass Roots conducted an important case study that showed how government decision-making is influenced by political negotiation and social considerations rather than just economic rationality. It challenged the traditional view of bureaucracy as a purely neutral and rational system. Instead, Selznick demonstrated that bureaucratic organizations are deeply political and shaped by social forces. His analysis of the Tennessee Valley Authority (TVA) revealed how local power dynamics, institutional culture, and informal relationships influence public administration.

The TVA, was a New Deal-era federal agency created in 1933 to promote regional economic development through infrastructure projects like dams and electricity generation. The TVA was originally designed as an apolitical, technocratic institution that would implement policy based on expertise rather than political considerations.

However, Selznick’s study showed that the TVA had to negotiate with local elites, businesses, and community groups to gain support for its programs. Rather than being a neutral bureaucracy, the TVA absorbed the interests and values of local stakeholders over time.
Political compromises often weakened the agency’s original mission of social reform and economic equality. For example, the TVA partnered with local conservative agricultural interests, even though these groups resisted social reforms that would have empowered poor farmers.

Selznick introduced the concept of “co-optation, which describes how bureaucratic organizations incorporate external groups to maintain stability and legitimacy. Instead of enforcing policies rigidly, agencies often have to adjust their goals to align with influential local actors. Co-optation helps agencies gain support and avoid resistance, but it can also dilute their original purpose. This explains why public organizations often fail to deliver radical change, even when they are designed to do so. For example, the TVA originally aimed to empower small farmers and promote land reform, but over time, it aligned itself with local business leaders and preserved existing power structures instead.

By embracing the principles of co-optation, governments can develop AI systems that serve the broader public interest and that development is guided by community engagement, transparency, and collaboration. AI development in government should involve active engagement with a wide range of stakeholders, including citizens, community groups, experts, and advocacy organizations. Co-optation can be used to address concerns and objections raised by external groups. By incorporating their feedback and making adjustments to AI systems, governments can mitigate potential opposition and build consensus.

Monopoly vs. Market Competition

Governments often hold a monopoly over essential services (e.g., national defense, law enforcement, public infrastructure) where competition is neither feasible nor desirable. Governments have broader responsibilities than businesses, encompassing national defense, social welfare, environmental protection, and infrastructure development. Technological changes, however, can change the dynamics of specific utilities. Telecommunications, for example, were primarily government-run facilities that worked to ensure universal service. To upgrade to the global Internet, however, these operations were largely deregulated or sold off to the private sector to invest in innovative new services. More recent discussions have pointed to “net neutrality” and even “cloud neutrality” to address the monopolization of services at the Internet’s edge, such as AI.

Leonard White (1926) in Introduction to the Study of Public Administration pointed out that government agencies do not face direct market competition, which affects incentives and operational efficiency. In contrast, businesses operate in a competitive market where consumer choice determines success. For example, the police department does not compete with private security firms in the way that Apple competes with Samsung in the smartphone market.

White also believed that public administration is the process of enforcing or fulfilling public policy. Since profit is not the primary goal, it’s crucial to define what constitutes “success” for AI systems in government. This might include citizen satisfaction, efficiency gains, improved outcomes, or reduced costs. By carefully considering the unique dynamics of government agencies and incorporating AI in a way that addresses the challenges of limited market feedback and different incentive structures, governments can leverage AI to create more effective, responsive, and citizen-centric services.

Legal and Ethical Constraints

Governments must operate under constitutional and legal constraints, ensuring adherence to democratic principles and human rights. Frank Goodnow (1900) in Politics and Administration argued that public administration is shaped by legal frameworks and public policy goals rather than market forces.

Public officials must follow strict ethical codes and conflict-of-interest regulations that go beyond corporate ethics policies. For example, a government agency should not arbitrarily cut services to boost its budget surplus, whereas a corporation can cut unprofitable product lines without legal repercussions.

Goodnow was one of the first scholars to formally separate “politics” from “administration”, arguing that politics involves the creation of laws and policies through elected representatives. Administration is the implementation of those laws and policies by bureaucratic agencies. Public administration should be neutral, professional, and guided by legal rules, rather than influenced by political pressures.

For example, Congress (politics) passes a law to regulate environmental pollution, and the Environmental Protection Agency (EPA) or the Federal Communications Commission (FCC) (administration) enforces and implements their laws and regulations through technical expertise and bureaucratic processes. Goodnow emphasized that public administration derives its legitimacy from legal and constitutional frameworks, not from market competition.

He argued that government agencies must operate within the rule of law, ensuring fairness, justice, and accountability. Laws define the scope of administrative power, unlike businesses that act based on profit incentives. Bureaucrats should be trained professionals who follow legal principles rather than respond to political or market forces. A tax agency must enforce tax laws uniformly, even if doing so is inefficient, whereas a private company can adjust its pricing strategies according to profit strategies.

Unlike businesses, which prioritize efficiency and profitability, Goodnow argued that government agencies serve the public interest. They provide services that markets might ignore (e.g., public health, education, law enforcement). Public agencies must prioritize equity, justice, and democratic values rather than cost-cutting. The effectiveness of government is measured not just by efficiency but by fairness and public trust. For example, governments fund public schools to ensure universal education, even if private schools might cater to specific family or community preferences.

By adhering to strict ethical principles and conflict-of-interest regulations, governments can ensure that AI is used in a way that builds trust, promotes fairness, and serves the public interest. It’s about creating AI systems that are not only effective but also ethical and accountable.

Service vs. Consumer Model

Citizens are not “customers” in the traditional sense because they do not “choose” whether to participate in government services (e.g., paying taxes, following laws). Harlan Cleveland, a distinguished diplomat and President of the University of Hawaii in his later years argued in his (1965) article “The Obligations of Public Power,” that public administration must ensure universal access to critical services, regardless of financial status. Businesses, on the other hand, serve paying customers and can exclude non-paying individuals from services. For example, a government hospital must treat all patients, including those who cannot afford to pay, whereas a private hospital can refuse service based on financial capacity.

His arguments focused on the ethical, political, and practical challenges faced by government officials in wielding public power. The ethical responsibility of public officials included holding power on behalf of the people, meaning they must act with integrity and accountability. Cleveland warned against abuse of power and the temptation for bureaucrats to act in self-interest rather than the public good. He stressed the need for ethical decision-making in government to prevent corruption and misuse of authority. For example, a government official responsible for allocating funds must ensure fairness and avoid favoritism, even when pressured by political influences.

Public administration should strive to be effective but must not sacrifice democratic values to pursue efficiency. He argued that bureaucratic decision-making should be transparent and participatory, ensuring citizens have a voice in government actions. Efficiency is important, but equity, justice, and citizen involvement are equally critical. For example, government programs should not cut social programs simply because they are expensive—public welfare must be prioritized alongside financial considerations.

Cleveland emphasized that public power must be answerable to multiple stakeholders, including the public (through elections and civic engagement), legislatures (through oversight and funding), and the courts (through legal constraints and judicial review). Unlike businesses, which are accountable mainly to shareholders, government agencies must navigate complex and often conflicting demands from different groups. For example, a public health agency must justify its policies to elected officials (who determine budgets) and citizens (who expect effective services).

Cleveland also pointed to the growing complexity of governance, a term he was one of the first to use. Government agencies were becoming more complex and specialized, requiring public administrators to manage technological advancements and expanding regulations as well as international relations and globalization. Cleveland worried that bureaucracies might become too rigid and disconnected from the people, creating a gap between government and citizens.

By keeping Cleveland’s principle at the forefront, governments can leverage AI to create a more just and equitable society where everyone has access to the services they need to thrive. It’s about using technology to empower individuals, reduce disparities, and ensure that everyone has the opportunity to reach their full potential.

As government agencies adopt AI and data-driven decision-making, they must ensure that technology serves human interests and does not lead to excessive bureaucracy or loss of personal agency. Cleveland called for adaptive, innovative leadership in public administration to keep up with social, political, and technological changes. He criticized government agencies that resist reform or fail to evolve with society’s needs. Public administrators must be proactive, responsive, and forward-thinking rather than merely following routine procedures. For example, climate change policies require public agencies to anticipate future risks, rather than simply reacting to disasters after they occur.

For Cleveland, public service was a moral obligation, not just a technical or managerial function. He believed that serving the public is an ethical duty, requiring commitment to justice, fairness, and the common good. Bureaucrats must see themselves as stewards of public trust, not just rule enforcers.

Harlan Cleveland’s emphasis on universal access to critical services, regardless of financial status, is a fundamental principle that must guide the design of AI mechanisms in government. Cleveland argued that public administration, unlike business, has a fundamental obligation to serve all citizens regardless of their ability to pay, and must balance efficiency with democratic values like equity, justice, and citizen participation. He stressed the ethical responsibility of public officials to act in the public interest, be accountable to multiple stakeholders, and adapt to the growing complexity of governance.

These principles are crucial for guiding AI development in government. AI systems should be designed provide universal access to critical services, overcoming barriers like financial constraints, location, and digital literacy. AI systems should avoid sacrificing democratic values in the pursuit of efficiency while maintaining transparency and accountability, allowing citizens to understand and participate in AI-driven decision-making. Ultimately, AI in government should be a tool for enhancing public service and promoting the common good, not just a means to increase efficiency.

Conclusion: The Unique Role of Government and the Implications of AI

Public administration scholars have consistently emphasized that government is not simply a business; it operates under different principles, constraints, and objectives. While efficiency is valuable, the government’s primary goal is to serve the public good, uphold democracy, and ensure fairness and justice, even at the cost of financial efficiency. The writings of Appleby, Cleveland, Waldo, Weber, and Wilson, continue to reinforce the fundamental distinction between governance and business management.

Drawing on the classics of public administration, we can start to specify some constraints and goals for artificial intelligence (AI) and develop a “smart” but ethical government that is efficient but also responsive to public concerns.

Possibilities and Oversight

– AI systems used in government should be transparent, with open access to the data and algorithms used to make decisions. This allows for public scrutiny and accountability.  
– Regular audits and oversight of AI systems are vital to ensure they function as intended and do not produce unintended consequences.  
– AI systems should be designed to protect the privacy of citizens and ensure that their data is used responsibly and ethically.  
– While AI can automate many tasks, human oversight is essential to ensure that AI systems are used in a way that aligns with ethical principles and societal values.  
– AI can make government information more accessible to citizens, providing clear and concise explanations of policies and programs.  
– AI can gather and analyze citizen feedback, providing valuable insights for policymaking and service delivery.  
– AI can facilitate participatory governance, enabling citizens to contribute to decision-making processes and shape public policy.  

Challenges and Considerations

– AI systems can perpetuate existing biases if not carefully designed and monitored. It’s essential to ensure that AI systems are fair and non-discriminatory.  
– The automation of specific government tasks may lead to job displacement. It’s important to develop strategies for workforce transition and retraining.
– Building and maintaining public trust in AI is crucial for its successful adoption in government. This implementation requires a commitment to transparency, explainability, and accountability in all AI-related processes and decisions.

By carefully considering these opportunities and challenges and drawing on the wisdom of the classics of public administration, we can start to harness the power of AI to create a “smart” but ethical government that serves the public interest and promotes the well-being of all citizens. In future posts, I plan to draw on subsequent generations of public administration practitioners and scholars who provide more critical perspectives of more complex government structures that have emerged in the last century. Women’s voices, such as Kathy Ferguson’s critique of bureaucracy and Stephanie Kelton’s critique of government budgeting, are extremely valuable perspectives going forward. AI is undoubtedly on the near horizon for government services, and it should be approached with the understanding that such systems are capable, but not guaranteed, of being designed for the public good.

Citation APA (7th Edition)

Pennings, A.J. (2025, Feb 09) AI and Government: Concerns from Classic Public Administration Writings. apennings.com https://apennings.com/digital-geography/ai-and-government-concerns-from-classic-public-administration-writings/

Bibliography

Appleby, P. H. (1945). Big Democracy. Alfred A. Knopf.
Cleveland, H. (1965). The Obligations of Public Power. Public Administration Review, 25(1), 1–6.
Ferguson, Kathy (1984). The Feminist Case Against Bureaucracy. Temple University Press.
Goodnow, F. J. (1900). Politics and Administration: A Study in Government. Macmillan.
Kelton, Stephanie. (2020) “The Deficit Myth: Modern Monetary Theory and the Birth of the People’s Economy.” PublicAffairs.
Selznick, P. (1949). TVA and the Grass Roots: A Study in the Sociology of Formal Organization. University of California Press.
Simon, H. A. (1947). Administrative Behavior: A Study of Decision-Making Processes in Administrative Organizations. Macmillan.
Waldo, D. (1948). The Administrative State: A Study of the Political Theory of American Public Administration. Ronald Press.
Weber, M. (1922). Economy and Society: An Outline of Interpretive Sociology (G. Roth & C. Wittich, Eds.). University of California Press.
White, L. D. (1926). Introduction to the Study of Public Administration. Macmillan.
Wilson, W. (1887). The Study of Administration. Political Science Quarterly, 2(2), 197–222. https://doi.org/10.2307/2139277

Note: Several AI requests were prompted and parsed for this post.

Share

© ALL RIGHTS RESERVED

---

Anthony J. Pennings, PhD is a Professor at the Department of Technology and Society, State University of New York, Korea and a Research Professor for Stony Brook University. He teaches AI policy and engineering economics. From 2002-2012 he taught digital economics and information systems management at New York University. He also taught in the Digital Media MBA at St. Edwards University in Austin, Texas, where he lives when not in Korea.

Related

Category: artificial intelligence, automatrix, crisis communications, democratic political economies, digital coordination, digital geography, digital media economics, disruption, enabling frameworks, enterprise systems, science and technology studies, smart new deal
Tags: bounded rationality > Dwight Waldo > Harland Cleveland > Herbert Simon > Max Weber > Paul Appleby > Philip Selznick > Tennessee Valley Authority (TVA) > Woodrow Wilson

Comments

Comments are closed.