Artificial Intelligence Takes Over Politics

Since the launch of ChatGPT — one of the most popular generative AI chatbots among users — artificial intelligence has become an integral part of people’s daily lives, assisting them in many routine activities. What was once known simply as an online chat tool has rapidly spread and been integrated into a wide range of products and services we use every day. For example, a simple Google search now often returns a synthesized summary called an “AI Overview.” Similarly, AI is built into the camera systems of next-generation smartphones, automatically enhancing photos to make them more visually appealing.

In recent weeks, Albania has taken a major step forward in integrating artificial intelligence into public services, opening a new chapter in national governance by introducing Diella, the world’s first government minister entirely created and managed by artificial intelligence. Unveiled by Prime Minister Edi Rama during the National Assembly of the Socialist Party, Diella has been tasked with overseeing all national public procurement procedures. Her primary objective is to tackle corruption and the misuse of public funds—issues that have plagued Tirana for decades and slowed the country’s EU accession process, which began in 2009.

Albania: From Digital Portal to Government Minister

Diella’s first role began in May 2022 as a virtual assistant for the e-Albania portal. At that time, it was essentially a digital tool that, using voice commands, helped citizens access over 95% of public services online. From the beginning, the avatar Diella, presented with feminine features and wearing an outfit inspired by traditional Albanian costumes, guided users through bureaucratic procedures and document requests, learning in real time from thousands of recorded interactions. The initiative’s success and high user satisfaction convinced the government to further expand her capabilities through a collaboration with Microsoft and the Azure OpenAI model.

In its new role at the ministry, the AI will go beyond handling administrative tasks and advising citizens; it will also autonomously organize and evaluate public procurement tenders. The handover from the ministries to Diella will take place gradually.

During the initial transition phase, the system will operate under human supervision, with ministry experts able to review decisions and step in if any anomalies arise. The ultimate goal, however, is to gradually give the AI greater decision-making autonomy, entrusting it entirely with the task of evaluating tenders impartially while ensuring full transparency by producing summary documents that public offices can easily access.

The Albanian government’s decision to assign public procurement oversight to an artificial intelligence system marks an unprecedented step in the country’s fight against corruption, which remains deeply entrenched in Albania. According to public opinion and several international observers, an impartial, “above politics” system could help dismantle the network of illegality that drains a significant share of public funds each year.

However, the initiative has also raised questions and concerns. Some commentators warn that this cutting-edge system could be vulnerable to hacking, algorithmic manipulation, and breaches of security protocols—risks that could compromise the integrity of the entire process, undermine the AI system’s impartiality, and ultimately jeopardize the credibility and reliability of the initiative.

Italy: AI Enters Parliament

The arrival of artificial intelligence on the Albanian political scene should not come as a major surprise to Italian citizens. In fact, AI is set to enter the chambers of the Italian Parliament through three chatbot platforms designed to assist lawmakers in drafting legislation and supporting them in their daily operations.

NORMA, developed from a proposal by the Politecnico di Milano and the Einaudi Institute, is the first tool to be made available to the Chamber of Deputies. It enables lawmakers to interact with the Chamber’s Legislation Observatory, identifying amendments and legislative texts using simple, natural language.

Meanwhile, the Alma Human AI inter-university consortium has developed the MSE system (Amendment Writing Machine), which actively assists parliamentarians in drafting and preparing amendments, streamlining the review process and accelerating the modification of legislative bills.

The final system, known as DEPUCHAT, was developed by Roma Tre University and Florence University to provide fast and reliable information on parliamentary activity. The system, the most ambitious and complex of the three, is intended for citizen use, allowing them to search for information and receive responses about individual parliamentarians, drawn from official profiles and records of their interventions in the chamber.

These three technological tools have been designed and developed over the past two years and, as Chamber Vice President Anna Ascani revealed, are the result of close collaboration between political staff, experts, and Big Tech representatives. Tools like these could prove essential in simplifying the drafting and review of legislation, which typically requires extensive research, while also helping prevent inconsistencies and duplication.

However, as Vice President Ascani explains, one of the key challenges remains preventing AI “hallucinations.” At the current stage of development, it is not uncommon for the algorithm, instead of admitting that it cannot answer a user’s question, to provide an approximate—or even entirely fabricated—response.

Despite its current limitations and the questions it raises, artificial intelligence is expected to assume greater responsibilities and enjoy increased autonomy as the technology continues to advance.

Translated by Iuliana Cindrea

Mondo Internazionale APS – All rights reserved ® 2025