By Denis Jukni

Artificial intelligence is rapidly moving from an experimental technology to a new infrastructure of the economy and public administration. Generative models, automation, robotics, cloud infrastructure, large-scale data processing and algorithmic systems are changing how companies produce, how institutions deliver services and how citizens receive information. This transformation opens opportunities to raise productivity, automate processes and improve public services. It also brings risks related to cybersecurity, privacy, algorithmic discrimination, disinformation, the labour market and the concentration of technological power.

For Albania, the question is no longer whether artificial intelligence will be used. It already is. The question is whether the country will manage to build the human, institutional, legal and technological capacities to use and control it.

From the race for algorithms to the race for infrastructure

The debate on artificial intelligence often focuses on models such as ChatGPT or Gemini. But the technological race is much broader. Developing AI requires considerable computing power, data centres, advanced processors, cloud services, large amounts of data, energy, capital and specialists.

The European Union makes this link explicit in its AI Continent Action Plan, built on computing infrastructure, data, skills, AI uptake and a trustworthy regulatory framework. In April 2026, the European Commission reported 19 “AI Factories” in Europe, with investment now extending to far larger “AI Gigafactories”.This adds a new dimension to the debate on digital sovereignty. A state can make extensive use of artificial intelligence while the models, processors, cloud infrastructure and platforms it relies on are controlled outside its territory.

For a small economy like Albania, replicating the entire global technology chain is not realistic. But it can build capacity to develop specialised applications, manage national data, audit algorithms, secure systems and avoid unnecessary dependence on a single technology supplier.

Innovation versus regulation

A central international debate concerns the line between innovation and control. Weak regulation can expose citizens and institutions to unsafe or discriminatory systems; poorly designed regulation can stifle innovation.

The European Union has opted for a risk-based approach through the Artificial Intelligence Act, Regulation (EU) 2024/1689. The Act entered into force in 2024, and significant parts of it already apply. Since 2 August 2026, the transparency obligations under Article 50, among other provisions, have been in effect. The logic is that not every use of AI carries the same risk: a system that recommends music cannot be treated like one that affects employment, education, public services or decisions with consequences for citizens.

Along the same lines, the OECD AI Principles call for transparency, explainability, safety, human oversight and accountability, including traceability throughout a system’s life cycle and the right of those adversely affected to understand and challenge an algorithmic outcome. UNESCO frames the issue directly in terms of human rights: dignity, privacy, fairness, transparency and human oversight must remain fundamental to the use of AI.

For Albania, these are no longer purely theoretical debates. In its Albania 2025 Report, the European Commission calls on the country to fully align with the AI Act and to put in place the structures needed to implement it.

Albania: rapid digitalisation, but a skills gap

Albania presents a striking contrast. According to the OECD’s Western Balkans Competitiveness Outlook 2024, the country has made considerable progress on digital society. Around 95% of public services were available online, and Albania ranked second in the Western Balkans on the digital society dimension.

Skills indicators, however, are far more problematic. Only 23.8% of individuals in Albania had basic or above-basic digital skills in OECD comparative data, compared with 34.8% in the Western Balkans and 53.9% in the EU. ICT specialists made up 2.6% of total employment, against 4.5% in the EU. Internet access, a digitalised administration and citizens’ ability to function in a digital environment are different things. Someone who uses a smartphone and social media every day is not necessarily able to recognise a sophisticated online scam, verify AI-generated material or understand an automated decision.

Albania compared with the Western Balkans and the EU

The table highlights a central contradiction: public administration has digitalised faster than the population has acquired digital skills. This human capital gap may limit Albania’s ability not only to use the technology, but also to develop, audit and control it.

Albanian businesses are entering the AI era

According to INSTAT data published on 21 September 2026, 12.4% of Albanian enterprises with 10 or more employees use artificial intelligence technologies, up from 9% a year earlier. Among large enterprises, the figure reaches 28.1%. Meanwhile, 35.3% of enterprises use cloud services and 34.5% employ ICT specialists. Uptake is growing quickly but remains below the EU level: according to Eurostat, around 20% of EU enterprises with at least 10 employees used AI in 2025.

The long-term problem runs deeper. The OECD finds that Albania performs above the Western Balkans average in most competitiveness dimensions, but below it in science, technology and innovation. Low investment in research and development, weak links between universities and business, and relatively modest scientific output limit the country’s capacity to move from being a user of technology to a creator of it.

Diella: from virtual assistant to public decision-making

Albania is also experimenting with AI in a far more sensitive field: state administration.

Diella was launched in January 2025 as the virtual assistant of the e-Albania portal. According to government figures, she had recorded 972,000 interactions by September of that year. She was then given the role of Minister of State for Artificial Intelligence, under the authority of the Prime Minister.

What matters is not the avatar or the title, but the functions to be delegated to the technology.

In 2026, the government announced that it is building an AI-supported electronic procurement system. According to the official description, it is meant to cover everything from market research and bid submission to contract monitoring. The government also states that Diella will be able to draw data from more than 40 public registers and that the process will be traceable and subject to manual review. These are the government’s objectives and claims about a system still under construction, not evidence that the problems of public procurement have been solved.

This is precisely where the question of algorithmic accountability arises. If an AI system influences a procurement procedure, who is responsible when it makes a mistake? Can it be established which data shaped the outcome? What happens if a state register contains inaccurate data? Can an economic operator obtain an explanation and challenge the result? Who has access to the system logs? Who audits the algorithms and the changes made to them?

These questions carry even more weight given the economic importance of public procurement and the European Commission’s repeated calls to strengthen competition, administrative capacity and professionalisation in the sector.

The use of AI in procurement may bring more automation and traceability, but technology does not by itself eliminate conflicts of interest, manipulation or corruption. If processes, data or control mechanisms are flawed, automation can reproduce those flaws on a larger scale.

Employment: transformation rather than sudden job losses

A major fear is that machines will replace workers.

A 2025 study by the International Labour Organization (ILO) estimates that one in four workers worldwide is in an occupation with some degree of exposure to generative AI. Only 3.3% of global employment falls into the highest exposure category. The ILO’s key conclusion is that, because most occupations still require human input, jobs are more likely to be transformed than eliminated outright.

In Albania, this could affect administration, finance, accounting, call centres, translation, media, legal services and other professions built largely on processing text and information. The challenge is not only how many jobs might be lost, but whether Albanian workers will acquire the skills to use AI as a productivity tool.

Risks: from cyberattacks to a crisis of trust

Artificial intelligence does not create one new category of risk so much as amplify existing ones. In cybersecurity, AI can help detect anomalies and attacks, but it can also be used to automate phishing and social engineering, produce convincing fraudulent content and analyse targets.

For Albania, this matters particularly after the major cyberattacks on government systems in 2022 and 2023. The OECD notes progress in aligning the cybersecurity framework with the EU acquis, but also challenges in ensuring the resilience of critical information infrastructure. Law No. 25/2024 “On Cybersecurity” and the National Cybersecurity Strategy 2025-2030 aim at alignment with European standards, including the NIS2 Directive.

Yet there is a paradox: the more state services, registers and algorithms are interconnected, the greater both the benefits of data exchange and the damage if they are compromised.

At the societal level, deepfakes, synthetic voices and the automated generation of text, images and video lower the cost of producing disinformation. The World Economic Forum’s Global Risks Report 2026 ranks misinformation and disinformation second and cyber insecurity sixth over a two-year horizon. “Adverse outcomes of AI” rises from 30th place over two years to fifth place over ten years. These reflect expert perceptions rather than statistical forecasts, but signal growing concern about the technology’s long-term consequences.

What should Albania do?

First, it needs a functioning AI governance framework aligned with the EU AI Act. Adopting rules is not enough: institutions need specialists able to assess algorithmic systems and enforce the rules.

Second, AI in public administration must come with mandatory algorithmic accountability: systems with significant impact need documentation, traceability, human oversight, security testing, complaint mechanisms and independent audits.

Third, digital skills should be treated as national infrastructure. An increasingly algorithmic administration cannot be built while many citizens lack basic digital skills.

Fourth, AI must be developed together with cybersecurity. Protecting digital identities, personal data, state registers and critical infrastructure must be built in from the start, not added afterwards.

Fifth, Albania should invest not only in buying technology but in the capacity to produce and evaluate it: universities, research, start-ups, open and high-quality data, Albanian language technology, and cooperation between academia, business and institutions.

Finally, technological sovereignty does not mean that Albania must produce every server, processor or large AI model itself. It means knowing whom it depends on, which data must not slip out of national control, how imported technology is audited, and how a supplier can be replaced without paralysing a public service.

Conclusion

Albania is not starting from zero. It has a largely digitalised administration, growing use of technology by businesses, a new cybersecurity framework and concrete AI experiments in public administration.

But the speed of adoption should not be mistaken for technological maturity. Gaps remain in digital skills, scientific research, innovation and institutional capacity.

The challenge for the coming years is therefore not simply to use more artificial intelligence. It is for Albania to have people who understand it, institutions that can control it, infrastructure that can protect it, businesses that can create value from it, and rules that ensure that, as algorithms take on an ever larger role in economic and public life, responsibility continues to rest with people.

Note: The indicators do not all refer to the same year. Digital skills, ICT specialists and the Digital Society index are taken from the OECD’s Western Balkans Competitiveness Outlook 2024. AI use by enterprises comes from INSTAT (Albania, 2026) and Eurostat (EU, 2025), so this comparison is indicative only.

About the author

Denis Jukni is a technical and cybersecurity expert, lecturer and contributor to the investigative platform Shteg.org (Network for Reporting on Organised Crime and Corruption in Albania). He studied Telematics in Graz, Austria (Universität Graz / TU Graz), and specialises in protecting digital infrastructure, cyber risk analysis and information systems security.

__________________

This text was produced within the framework of the MediaLit project, implemented by the Centre for Civic Education (CGO) in partnership with the Helsinki Committee for Human Rights in Serbia, the Helsinki Committee for Human Rights in North Macedonia, the Tirana Centre for Journalistic Excellence in Albania, and the Atlantic Initiative in Bosnia and Herzegovina. The project is funded by the European Union and co-financed by the Ministry of Regional and Investment Development and Cooperation with NGOs of the Government of Montenegro. The content of this text is the sole responsibility of the author and does not necessarily reflect the views of CGO, the partner organisations, the European Union, or the Ministry.

The post Technological Development in the Age of Artificial Intelligence: Global Debates, Risks and Prospects for Albania appeared first on Tirana Times.

Original post Here