Albania has taken an unprecedented step in governance, appointing an artificial intelligence–generated bot as its newest cabinet member, responsible for public procurement.
Prime Minister Edi Rama, who is preparing to begin his fourth term in office, announced on Thursday that the AI official, named Diella — which translates as “sun” in Albanian — will oversee the awarding of government tenders.
“Diella is the first cabinet member who isn’t physically present, but is virtually created by AI,” Rama said in a speech unveiling his new cabinet.
He declared that the system would ensure Albania becomes “a country where public tenders are 100% free of corruption.”
A Bold Experiment Against Graft
Public procurement has long been plagued by corruption scandals in Albania, with lucrative contracts often cited as vehicles for bribery and influence-peddling. Analysts say this has helped turn the Balkan state into a hub for international criminal networks seeking to launder drug and weapons trafficking proceeds.
The perception of endemic graft has also hampered Albania’s bid to join the European Union, which Rama has vowed to achieve by 2030 — a timeline experts call highly ambitious.
By installing Diella as a gatekeeper, Rama is betting that AI can succeed where human officials have repeatedly failed, cutting off avenues for bribery, threats, or favor-trading in public tenders.
How Diella Works
Details remain sparse about what kind of human oversight will accompany Diella’s decision-making or what safeguards exist to prevent manipulation of the system itself. The government did not address these risks, raising questions about transparency and accountability.
Diella was first introduced earlier this year as a virtual assistant on the e-Albania platform, where she helped citizens and businesses obtain official documents. Dressed in traditional Albanian attire, she answers voice commands and issues electronically stamped paperwork, designed to reduce bureaucratic delays and limit petty corruption.
Public Skepticism
The public response to her promotion has been divided. On Facebook, one user by Reuters quipped: “Even Diella will be corrupted in Albania.” Another was more resigned: “Stealing will continue and Diella will be blamed.”
Their skepticism reflects broader doubts about whether technology alone can uproot systemic corruption that has persisted for decades and penetrated the highest levels of government.
The newly elected parliament, which came to power in May, will convene on Friday. It is not yet clear whether Rama’s government lineup, including Diella’s unprecedented “appointment,” will be formally voted on the same day.
Analysts have noted three possible outcomes of the AI.
Best Outcome: transparency, integrity, and an EU boost
Weighing in, analysts believe that if Diella is implemented with rigorous, transparent safeguards, the result could be a meaningful reduction in rent-seeking around public tenders. That would require the government to publish the procurement algorithms, logging and immutable audit trails, independent oversight (including human review panels and judicial recourse), and strong cybersecurity to prevent manipulation.
In this scenario, automated tender evaluation reduces discretionary decision points, limits opportunities for bribes and coercion, and speeds contract awards. Citizens and international monitors see clearer, evidence-based procurement records; corruption complaints fall; and Albania’s standing with EU accession officials improves.
Diella’s role on e-Albania as a document issuer — already reducing face-to-face friction — becomes a template for digitized services that shrink petty graft across the bureaucracy. The success would not only modernize procurement workflows but also send a political signal that Rama’s administration can tackle long-standing governance issues.
Intermediate outcome – limited gains, political theatre, and partial reform
A more likely middle path is partial improvement coupled with persistent gaps. Diella’s automated processes may cut some avenues for low-level corruption (for example, paperwork irregularities and administrative delays), and pilots on select categories of procurement could run smoothly. But without full disclosure of decision rules, continuing political influence over contract scopes and specifications could reintroduce human discretion earlier in the pipeline. Corrupt actors might shift tactics rather than disappear — for instance, by gaming pre-qualification requirements, influencing the terms of reference, or targeting subcontracting and implementation phases that remain under human control.
Public confidence would rise modestly in areas where the bot demonstrably expedited procurement, but scepticism would remain where high-value contracts and strategic projects are concerned. EU negotiators may note improvements in the process but demand deeper legal and institutional reforms before treating Diella as proof of systemic change.
Worst outcome — façade, manipulation, and reputational damage
If the government deploys Diella without meaningful checks, the bot risks becoming a scapegoat or, worse, a tool for automated graft. The absence of transparent oversight could allow bad actors to manipulate input data, alter evaluation criteria, or capture the system through insiders who control configuration or the human interfaces that feed the AI. Algorithmic bias could systematically disadvantage certain bidders, favoring politically connected firms while cloaking decisions behind technical complexity.
Public distrust would likely deepen if high-profile tenders awarded by Diella are later linked to poor delivery, inflated costs, or conflicts of interest. International partners and the EU could view the initiative as cosmetic, intensifying scepticism about Albania’s reform commitments and making accession prospects already ambitious for 2030 harder to substantiate.
Overall, Diella’s fate is expected to depend on explicit answers to a handful of questions: Who controls the model’s source code and training data? Will procurement criteria and scoring rubrics be published before tenders open? What independent, human-led appeals process will exist for losing bidders? How will the government secure the system against tampering, and what audit trails will be made available to civil society, media, and international monitors?
Why this matters beyond Albania
Global interest in algorithmic procurement is rising; countries from developed democracies to emerging economies are experimenting with automation to reduce corruption. Albania’s Diella will be watched not only for what it accomplishes domestically but for whether a small state can use AI to confront a problem that has resisted political remedies. A successful model could be exported; a failed one could become a cautionary tale about outsourcing governance to opaque systems.
