Carsten Maschmeyer's naive wishful thinking? AI revolution in German administration: The core organizational problem – Image: Xpert.Digital
What radical demand has Carsten Maschmeyer made regarding the German administration? Vision or dangerous illusion?
Enough with the bureaucratic madness? The truth behind Germany's core organizational problem
Can artificial intelligence really save our government offices?
Carsten Maschmeyer, known as an investor from the TV show "Die Höhle der Löwen" (The Lion's Den), made a far-reaching demand in a January 2026 interview with the Neue Osnabrücker Zeitung: German public administration must be replaced by artificial intelligence, virtually entirely. His argument follows a technocratic logic aimed at maximizing efficiency. Maschmeyer promises that with comprehensive AI implementation, decisions could be made within seconds. In his view, this would finally put Germany back at the forefront of innovative administrative structures worldwide.
The investor justifies his proposal by citing the current state of German authorities, which he characterizes as too slow, too expensive, and outdated. He particularly criticizes the lack of creativity in many administrative tasks: Checking whether all the boxes are correctly ticked on an application for a new identity card, he argues, requires no human creativity. In his view, permits, grant applications, and administrative acts could easily be processed by a computer. Only exceptional cases should continue to be handled by humans.
The economic implications of this vision are considerable. Maschmeyer argues that the drastic staff reductions in government agencies would also significantly reduce pension liabilities in the long term. He claims his primary goal is not layoffs, but rather speed and increased efficiency. As a concrete example, he cites a major German city where, despite a 30 percent decrease in building permit applications, processing times have doubled. This inefficiency, he says, is incomprehensible to everyone.
Related to this:
What fundamental problem lies behind German administrative bureaucracy?
The root cause of the problems in the German administration lies not in the malicious actions of individual civil servants, but in the institutional logic of the system itself. Every institution develops structural self-interests that work against simplification and the reduction of bureaucracy. This phenomenon is well documented in administrative science and describes the tendency of organizations toward self-preservation and growth.
Institutional self-interest manifests itself on several levels. First, the administration as a whole has an interest in maintaining or expanding its significance and resources. Second, individual departments and employees benefit from complex structures that make their specific expertise indispensable. Third, legal requirements and processes create path dependencies that are difficult to break. The administration can primarily only react, not act, as it is bound by the law.
An often overlooked aspect is the role of external consultants in this system. Consulting firms have developed a business model that profits from the complexity of bureaucracy and is geared towards this in the long term. The more complicated the administrative structures, the greater the need for consulting. These consultants therefore have no interest in radical simplification, but rather earn their living by optimizing existing complex systems. This creates a perverse incentive system in which those who are supposed to offer solutions profit from the existence of the problem.
The lack of corrective measures further exacerbates the situation. Political leaders are often unable or unwilling to implement structural reforms because these encounter considerable resistance. Behind almost every administrative structure and legal regulation lie special interests that benefit from the status quo. This leads to a kind of double standard in public discourse: everyone calls for deregulation and reduced bureaucracy, but when it comes down to specifics, those affected defend their existing regulations.
Related to this:
What are the actual costs of German bureaucracy?
The scale of bureaucracy in Germany is immense. According to current figures from the Federal Statistical Office from January 2026, the bureaucratic costs for German companies due to reporting obligations alone amount to €62.5 billion per year. This figure has decreased slightly compared to the previous year, but remains at an extremely high level. The number of reporting obligations was reduced from 12,390 in January 2025 to 12,364 – only a marginal improvement.
A more comprehensive study by the ifo Institute arrives at even more dramatic conclusions. Researchers calculated both the direct and indirect costs of bureaucracy and concluded that Germany loses up to €146 billion in economic output annually due to excessive red tape. This estimate takes into account not only the direct costs of compliance but also the indirect costs and opportunity costs resulting from delayed projects, tied-up capital, and legal uncertainty.
The burden is particularly severe for small and medium-sized enterprises (SMEs). A study by the Association of German Chambers of Industry and Commerce (DIHK) on the hospitality industry reveals that companies in this sector face 125 legal obligations, 43 percent of which are industry-specific. Alarmingly, 40 to 70 percent of these obligations are not related to the actual business processes but merely serve bureaucratic requirements. SMEs in the hospitality industry have to spend an average of 2.5 percent of their annual revenue on red tape, which equates to between €12,000 and €60,000 per year. Many business owners work an average of 14 hours of overtime per week just to comply with government regulations.
In international comparisons, Germany performs particularly poorly. According to OECD data, Germany regularly ranks in the top third of industrialized countries for the duration of approval and planning procedures, especially for construction and infrastructure projects. The time spent on tax return compliance in Germany, at 218 hours per year, is almost twice as high as in Sweden, where it is 122 hours.
Why is Maschmeyer's demand considered simplistic?
Yes, it's a striking image, but it's a way to get the public moving! The media needs to do more with it!
Maschmeyer's statements may appear deceptively simple at first glance, but upon closer examination they prove to be unrealistic and dangerously simplistic. The demand for a near-complete replacement of public administration by AI ignores fundamental legal, technical, and social realities. It is not a well-thought-out reform strategy, but rather wishful thinking that fails to grasp the complexity of modern government.
Maschmeyer's presentation completely ignores the technical limitations. While AI systems can efficiently handle structured tasks with clear rules, they quickly reach their limits when faced with complex case-by-case decisions, discretionary leeway, and the balancing of conflicting interests. Public administration is characterized precisely by such complex balancing decisions, which cannot simply be translated into algorithms. AI can, under certain circumstances, classify texts, review documents, pre-sort cases, and provide indications of inconsistencies, but the responsibility for legally binding decisions cannot simply be delegated to machines.
The legal hurdles are considerable. Public administration is subject to strict data protection and regulatory requirements. The General Data Protection Regulation (GDPR) sets high standards for automated decisions, especially when these have legal effects or significantly affect data subjects. According to Article 22 of the GDPR, data subjects generally have the right not to be subject to a decision based solely on automated processing. This applies particularly to decisions of significant importance.
The lack of transparency in many AI systems contradicts the principles of the rule of law. Administrative decisions must be comprehensible and justifiable. However, AI systems often operate as a "black box," whose decision-making processes are not transparent. This leads to significant problems with legal oversight and the protection of fundamental rights. Incorrect AI results, so-called hallucinations, and potential bias in results due to bias in the training data pose additional risks.
What specific challenges are preventing the widespread use of AI in public administration?
The list of obstacles to the widespread use of AI in public administration is long and fundamental. Foremost among them is the lack of regulatory clarity. Many AI projects fail as early as the planning phase because it is unclear how legal requirements can be met. Uncertainty surrounding the GDPR and other regulations often leads public authorities to be hesitant in implementing AI solutions. Standardized guidelines and experience in handling sensitive data are frequently lacking. Public authorities have legitimate concerns about data protection breaches, which can result in hefty fines of up to €30 million or six percent of global annual turnover.
Data quality presents another key challenge. AI systems are only as good as the data they are trained on. In public administration, data is often fragmented, non-standardized, and distributed across various systems. Germany's federal structure, with its distributed responsibilities at the federal, state, and local levels, significantly exacerbates this problem. Inconsistent regulations at different levels further complicate the implementation of AI projects.
The shortage of skilled workers and a lack of technical expertise within public authorities are significantly hindering implementation. Many municipalities lack both the financial resources and the personnel capacity to implement AI projects. According to data from the Federal Statistical Office, one-third of public sector employees will retire by 2035. This will lead to a dramatic loss of knowledge and exacerbate the staff shortage. At the same time, there is a lack of young talent, particularly with digital and AI skills.
Political and organizational inertia should also not be underestimated. Administrative reforms are traditionally difficult to implement in Germany. The fragmented state architecture and the highly fragmented federal responsibilities do not provide fertile ground for goal-oriented administrative reform. There is no coherent administrative policy in Germany that unites the different levels of government. Parliamentary interest in administrative policy has so far been insufficiently developed.
Our EU and German expertise in business development, sales and marketing
Our EU and German expertise in business development, sales and marketing - Image: Xpert.Digital
Industry focus areas: B2B, digitalization (from AI to XR), mechanical engineering, logistics, renewable energies and industry
More information here:
A thematic hub offering insights and expertise:
- Knowledge platform covering global and regional economies, innovation and industry-specific trends
- A collection of analyses, insights, and background information from our key areas of focus
- A place for expertise and information on current developments in business and technology
- A hub for companies seeking information on markets, digitalization, and industry innovations
The real reason for bureaucracy: Why AI alone won't save our government offices
What risks are associated with rushing to automate administration?
The dangers of overly rapid and comprehensive automation are manifold and potentially serious. Foremost among these is the risk of systematic discrimination. AI systems can adopt biases and distortions from their training data and reproduce them on a large scale. For example, if historical administrative decisions containing discriminatory patterns are used as training data, the AI perpetuates this discrimination. A particularly problematic aspect is that such systematic errors are often difficult to detect and can structurally disadvantage certain population groups.
The loss of individual justice is another key problem. Administrative decisions often require consideration of individual circumstances, hardship cases, and special situations. Standardized algorithmic processing cannot do justice to these specificities. Reducing complex social realities to binary decisions leads to injustice and social hardship. This contradicts the principle of individual justice, which is a core principle of the rule of law.
The lack of transparency in automated decisions undermines trust in state institutions. When citizens don't understand why an authority has made a particular decision and no human contact person is available, this leads to frustration and alienation. The feeling of being at the mercy of an opaque machine can be a threat to democracy. Public administration loses its human dimension and becomes an impersonal, technocratic apparatus.
Dependence on technology providers creates new risks. When critical administrative functions depend on proprietary AI systems, the state becomes reliant on private companies. This affects both technical dependence and data sovereignty. System failures, security vulnerabilities, or financial problems of the providers can then directly impair the state's ability to act. The increasing integration of AI into everyday life leads to growing dependence, often accompanied by a lack of understanding of AI's limitations.
Related to this:
What does previous experience with AI projects in public administration show?
The reality of AI implementation in public administration is sobering. While AI is already being used in many private sector companies to optimize processes, its application in public administration is often still in its infancy. The majority of organizations are in an experimental or exploratory phase, where fundamental questions about technology, data protection, and application possibilities still need to be answered.
Many administrative processes are highly standardized and offer little room for innovative approaches. While companies can react agilely to new technologies, public authorities are often bound by stricter legal frameworks that slow down change. Although studies show that, theoretically, up to 82 percent of administrative staff could be relieved of tasks by AI technologies, successful AI projects in public administration are rare. The gap between theoretical potential and practical implementation is enormous.
The few existing pilot projects often remain isolated solutions that cannot be scaled. They are hampered by data protection regulations, a lack of IT standards, and political caution. One example of an ambitious approach is the KärntenGPT project in Austria, which aims to demonstrate whether an administration can remain functional despite massive staff reductions through technological substitution. This project is considered a first stress test and demonstrates that the vision of a largely automated administration is at least partially technically feasible, even if the long-term effects are still unclear.
The shortcomings of previous digitization efforts are structural in nature. Many projects have shown that it wasn't the individual tools and measures that were inadequate, but rather the overall structure that was flawed. Without a comprehensive process analysis, clear data standards, and transparent justification mechanisms, new contradictions, legal proceedings, and control loops arise, negating the hoped-for acceleration. AI thus becomes a cost driver instead of a source of relief.
What are some realistic use cases for AI in public administration?
Despite all the challenges, there are indeed sensible and practical applications for AI in public administration that are far removed from Maschmeyer's maximum-maximum demand. The most promising approaches follow a phased approach: first, assistance systems and automated preliminary reviews; then, semi-automated decisions with human oversight; and only lastly, further automation in clearly defined areas.
Standardized chatbots and digital assistants to support citizen services offer good starting points. They can provide assistance with frequently asked questions about waste collection, forms, or opening hours. Furthermore, open-source solutions or cooperation with municipal IT service providers allow for the cost-effective implementation of document classification, appointment management, or simple automated processes such as application processing. These applications offer genuine added value without legal or ethical risks.
Extracting documents, verifying the plausibility of information, assigning responsibilities, and prioritizing cases based on risk are further sensible applications. AI can truly provide relief here if it is designed as an assistance system that prepares decisions and flags deviations, rather than making final decisions itself. The ultimate responsibility and decision-making power must remain with humans, who can critically review the AI's suggestions and overrule them if necessary.
In some areas, AI can also help in recognizing patterns and anomalies, such as in fraud detection or in identifying cases that require special attention. It is important that AI is not understood as a replacement for human expertise, but rather as a support tool. The systems must be transparent, traceable, and regularly checked for bias. A systematic review of opportunities and risks emphasizes that impact measurement, governance, and risk management often determine success or failure.
What needs to happen to achieve genuine bureaucracy reduction?
The key to genuine administrative reform lies not in a complete technological overhaul, but in a fundamental critique of tasks and process optimization. Before AI is deployed, it is essential to examine which regulations, verification processes, and documentation requirements are still meaningful and necessary. Many regulations have accumulated over decades without ever being systematically reviewed for their necessity. A radical critique of tasks, asking which government services must actually be provided and how these can be organized most efficiently, is the fundamental prerequisite.
Simplifying and standardizing processes must take precedence over automation. Simply replicating inefficient analog processes digitally is pointless. Instead, procedures must be fundamentally rethought and designed from the user's perspective. The "once-only" principle—data only needs to be transmitted to the government once and can then be shared internally—is crucial here. However, this requires overcoming siloed thinking and federal infighting.
The political will to implement measures is crucial. Bureaucracy reduction fails not due to a lack of understanding of the problems, but rather due to a lack of political will to overcome institutional resistance. Binding mechanisms are needed, such as automatic regulation reduction, where for every new regulation, an old one must be repealed. Introducing expiration dates for regulations, after which they automatically expire if not actively renewed, could help halt the continuous increase in regulatory density.
Creating societal pressure through information is also essential. Citizens must understand how much bureaucracy stifles innovation, destroys jobs, and erodes prosperity. Only when broad public awareness of the urgency of genuine reforms arises will politicians be compelled to act. Transparency regarding the actual costs and impacts of regulations is crucial. Instead of hoping for technological miracle solutions, we must wake up and build societal pressure on this system through continuous information dissemination.
What should a responsible digitalization strategy look like?
A sensible strategy for modernizing public administration must adhere to several principles. First, it must put people at the center, not technology. The use of AI must not be an end in itself, but must be aligned with clear objectives, legal requirements, and ethical principles. Only through a smart combination of technological innovation, legal framework, and organizational responsibility can AI contribute to strengthening public administration.
Secondly, a realistic assessment of AI's capabilities and limitations is needed. AI is a tool, not a panacea. Systems must be designed with sufficient transparency in their operation so that users can appropriately interpret and utilize the results. This is particularly important because AI use in public administration is considered a high-risk application and is subject to strict regulations. Affected employees must be trained to understand the limitations and capabilities of AI systems so they can regularly review expenditures.
Thirdly, a step-by-step, iterative approach is necessary. Instead of large-scale transformation projects, pilot projects should first be conducted in clearly defined areas, evaluated, and scaled if successful. These projects must be designed from the outset to be legally sound, transparent, and involve all stakeholders. Success measurement must encompass not only efficiency gains but also the quality of decisions, citizen satisfaction, and legal certainty.
Fourth, digital sovereignty must be preserved. The state must not become entirely dependent on private technology providers. This requires investment in its own expertise, the use of open-source solutions where possible, and clear contractual agreements on data protection and control over critical systems. The administration must remain capable of fulfilling its core functions even without AI to ensure resilience.
