Diella AI
Albania becomes the world’s first country to appoint an Artificial Intelligence (AI) made minister, Diella, developed in partnership with Microsoft to help fight corruption.
Following a presidential decree in September 2025, Diella was formally appointed as Minister of State for Artificial Intelligence. Procurement responsibilities are planned to be transferred gradually to the system to reduce political influence in tender procedures in the country.
This is part of Prime Minister Edi Rama’s anti corruption reforms and measures intended to curb corruption and align Albania with the European Union accession requirements.
Diella joins Humphrey and Albert who also play significant roles in their respective governments. The Humphery suite was developed by the UK government’s Incubator for Artificial Intelligence with a mandate to improve public services and government processes.
Albert API on the other hand helps public service agents respond to online requests, making correct use of state information, producing verifiable content and adopting the tone of the administration.
Kenya has a proactive stance on technology and AI, aiming to become Africa’s leading hub for AI research, innovation, and ethical development as primarily guided by by its comprehensive National AI Strategy 2025-2030.
With various Ministries in Kenya facing numerous scandals, including the Health Ministry’s Social Health Authority (SHA) which has been rocked by large-scale fraud allegations with Health Cabinet Secretary Aden Duale admitting that 35 hospitals are accused of stealing over KSh 104.5 billion from the health fund through fraudulent claims.
Separately, there is an ongoing investigation into a Sh443.6 million oxygen plant tender for Kenyatta National Hospital involving allegations of falsified documents and lack of due diligence, and reports of “ghost hospitals” being paid millions by the SHA.
If Kenya gave mandate to an AI powered Minister to head one of the ministries, say that of Health, conceptually, the following could happen.
AI Minister can reduce fraud in SHA by ensuring hospitals are paid correctly and timely by automatically handling the medical claims, which reduces errors and speeds up the process. The technology can also spot suspicious billing patterns and check accuracy of medical codes to avoid payment mistakes. Also, AI uses past data to predict future costs, allowing SHA to manage its budget and plan its finances more effectively.
An AI Minster could use predictive analytics to forecast disease outbreaks, manage drug supplies (preventing stock-outs), proactively allocate resources to regions in need, which would be based purely on data trends since it is driven by processing vast health datasets to suggest evidence-based policies and decisions.
Since AI processes massive amounts of information rapidly, it could enable faster policy development and implementation of initiatives like mHealth and AI diagnostics.
AI is immune to personal gain or political pressure, unless trained on biased data sets, meaning that when bias is eliminated in training, it is capable of following programmed ethics and logic which will do away with political influence and corruption in accessing medical tenders.
Unlike humans who are limited by physical and emotional capacities like fatigue and emotional exhaustion, AI powered Minister can operate 24/7 and capable of managing large volumes of administrative tasks and monitoring systems continuously.
However, AI lacks human intuition, empathy, and cultural sensitivity and requires new, robust legal and ethical frameworks to define accountability in cases of error or harm, which are currently lacking in Kenya.
Ultimately, an AI “minister” in Kenya would function less as a replacement for human leadership and more as a powerful, data-driven advisory and implementation tool just like the SERA.ai tool by the ministry of Health that is designed to shorten the country’s health policy formulation process which is mostly a lengthy bureaucratic processes and poses difficulties in consolidating research and evidence.
