The Future of AI Governance: Diella, Albania's AI Minister
By Mejreme Asllani | 26 October 2025
Summary
Albania has appointed “Diella”, an artificial intelligence (AI) minister, who has been granted executive authority over public procurement to combat corruption and improve transparency.
This high-stakes experiment presents immediate constitutional challenges and introduces significant operational risks, including algorithmic bias and cyber threats.
Whether successful or not, Diella’s long-term impact will likely reshape standards for AI in public administration and encourage the development of new legal and ethical frameworks to govern autonomous political actors.
Context
In September 2025, Albania made global headlines by appointing ‘Diella’, an AI-powered Minister of State for Public Procurement. The government, led by Prime Minister Edi Rama, who was elected for a fourth term in May 2025, introduced ‘Diella‘ as a non-human minister within the cabinet. Diella has been entrusted with the critical and highly sensitive portfolio of public procurement. Its primary mandate is to oversee and manage government tender processes and to ensure they remain free from corruption, a deeply entrenched issue in Albania’s political landscape. This appointment is part of Albania’s broader strategy to utilise technology in its anti-corruption efforts, which aligns with its goal to join the European Union (EU) by 2030.
The system was developed in a strategic partnership with Microsoft. Diella initially functioned as a public-facing virtual assistant on the ‘e-Albania’ platform that offers online services for citizens and businesses. Following a probationary period, during which it helped issue over 36,600 digital documents and provide nearly 1,000 services, Diella was elevated to a cabinet-level role, introducing a bold experiment in AI governance. The initiative seeks to provide a technological solution for a deeply entrenched human and political problem, making Albania the first nation to formally embed an AI within its cabinet structure as a designated, autonomous political actor. However, the appointment has also spurred controversy and debate regarding constitutional legality, ethical governance, and the transparency of AI decision-making.
Constitutional and Political Challenges
Diella’s appointment challenges conventional definitions of governance by introducing a non-human actor with decision-making powers at the ministerial level. While symbolically asserted as a step toward combating corruption through automated and unbiased oversight, critics argue that the move could conceal entrenched political interests by masking them as technological innovation, raising concerns about the true motivations behind this bold experiment. Albania’s opposition parties have voiced strong dissent, arguing that Diella’s role might mask corruption instead of resolving it, a stance that has sparked uproar in parliament. Additionally, constitutional questions arise as Albanian law mandates ministers be natural persons, rendering Diella’s official legal status ambiguous.
This ambiguity necessitates human validation of AI decisions, limiting its autonomy but highlighting tensions between legal frameworks and technology. For the Albanian government, it is a high-reward, high-risk strategy. Success could significantly boost its domestic legitimacy, unlock substantial economic efficiencies, and accelerate its EU accession timeline. However, the potential for failure is equally high. If Diella is perceived as a mere “publicity stunt”, or if its decisions are successfully challenged on constitutional grounds, it could severely damage public trust not only in the government but also in the very concept of technological solutions for societal problems. Critics suggest this move risks being remembered more as propaganda than as policy, a political performance carefully staged to claim a 'world first' while sidestepping the fact that there is no clear constitutional basis for an AI to hold ministerial office.
EU Integration and Regulatory Alignment
The initiative also has significant implications for Albania’s EU integration process. As a country aiming for accession by 2030, Albania must demonstrate compliance with the full body of EU law, including critical regulations governing technology and data. Consequently, this experiment will be closely monitored by Brussels as a test of the country’s commitment to aligning with EU digital governance and transparency standards. Moreover, to gain credibility, Albania must ensure that Diella’s AI-powered activities adhere to the EU AI Act, the General Data Protection Regulation (GDPR), the Digital Market Act (DMA) and broader EU standards and regulations, demonstrating not only technological advancement but also ethical compliance and institutional accountability.
Cybersecurity and National Security
From a security perspective, centralising all procurement authority within a single digital entity creates an extraordinarily high-value target for state and non-state cyberattacks. Since 2022, Albania has faced multiple cyberattacks attributed to Iran-affiliated “Homeland Justice” groups that disrupted government networks, followed by continued assaults during 2023–2024 targeting the Parliament, the national statistics agency (INSTAT), telecom providers, and media institutions, collectively exposing systemic vulnerabilities across its digital ecosystem.
In response, Albania has accelerated comprehensive cybersecurity reforms through the National Cybersecurity Strategy (2020–2025), the new Law on Cybersecurity (Law No. 25/2024) aligned with EU directives, and the strengthening of its National Cyber Security Authority (AKSK) and CSIRT structures. These measures are complemented by strategic partnerships with the US, NATO, and the EU, which focus on capacity-building, incident response coordination, and resilience training to strengthen national defences against persistent state and non-state-sponsored cyber threats.
When it comes to Diella, the threat is not merely a data breach or Distributed Denial of Service (DDoS) attacks, but rather sophisticated adversarial attacks like ‘data poisoning’, where malicious actors subtly manipulate the training data over time to make the AI favour certain outcomes, thereby exploiting the system for foreign influence or economic espionage. As the AI minister oversees large-scale governmental contracts and financial flows, securing its infrastructure against cyber threats is paramount.
The ‘GovTech’ Precedent
Beyond Albania’s specific challenges, governments worldwide are increasingly integrating AI into public institutions to increase efficiency and transparency and reduce bureaucracy and decision-making processes. However, Diella represents a significant leap beyond simple integration, establishing a new frontier for digital governance by positioning an AI as an executive, non-human actor within the government. This experiment could open up a significant new market for AI in public administration in the region, but it also places immense reputational risk on corporate partners like Microsoft. A high-profile failure could lead to a public backlash against the use of AI in sensitive government roles.
Forecast
Short-term (Now - 3 months)
It is highly likely that the Albanian opposition will initiate formal constitutional court challenges questioning Diella's legal status and decision-making authority.
It is also highly likely that experts in the field will closely monitor Diella's initial actions, making Albania a focal point in the Western Balkans discourse on AI ethics, digital sovereignty, and algorithmic governance.
It is highly likely that Diella's precise role and decision-making criteria throughout the procurement process will face intense scrutiny, leading to demands for greater transparency and accountability.
Medium-term (3-12 months)
It is likely that a controversial, high-value government tender decision made by Diella, potentially involving a major infrastructure project or influential domestic company, will serve as a critical test of its legitimacy, transparency, and ability to resist political and corporate pressures.
It is also likely that the European Commission will include formal guidance or policy recommendations on AI deployment in core government functions within its progress report on Albania, setting benchmarks for candidate nations.
Long-term (>1 year)
It is highly likely that the evolving need for independent verification of Diella’s fairness and effectiveness will foster the growth of a dedicated professional and academic sector focused on algorithmic auditing for public-sector AI applications.
It is highly likely that the critical risks of AI sovereignty and data integrity will become central to the policy debate, making them paramount issues for the system's long-term viability and credibility.
If this experiment becomes successful, it is likely that other nations may adopt similar virtual or AI models with such executive roles.
