Albania just appointed an AI agent, called Diella, as minister for public procurement. Its main tasks are to reduce corruption, increase transparency, and encourage innovation in this important area. Before this role, Diella worked as a virtual assistant on the e-Albania public service platform.
Legally, this appointment is just symbolic because, according to the Albanian Constitution, ministers must be citizens who are at least 18 years old. However, this situation blurs the lines between technology and actual government roles. We spoke with Volodymyr Getmanskyi to explore the implications of having an AI-powered minister.
Background & experience:
- Over 15 years of practical experience in advanced data analysis and modelling. Currently manages a large AI team while providing presales and delivery support for complex AI implementations.
- Technical expertise encompasses the full spectrum of AI technologies relevant to government applications: NLP/AI agents for document/information processing and assistance, compliance checks, computer vision for security and monitoring systems, AI-based RPA for process automation, and predictive modelling for policy planning and resource optimisation.
What makes Diella's appointment to a government ministry so significant?
Volodymyr Getmanskyi: First of all, this is a great and first illustrative case where AI agents are publicly involved in such an important national function. As stated, Diella initially helped with government services usage on the website, but after half a year, it was promoted to a member of the ministry. Also, it was developed with support from current and former OpenAI experts, removing a number of questions related to the AI agent's customisation and quality-based risks.
Besides that, please take into account the functions which are delegated to the AI agent. It is about procurement evaluation, where typically there should be some quantitative underlying comparison. It significantly decreases AI agent risks, because it is mostly a deterministic comparison (price higher/lower, etc.), especially if there are defined criteria, which is the most common practice in most government/public tenders. I assume that other qualitative comparisons are made under strict instructions with reasoning, thus the agent can be more robust, transparent, and explainable.
Are there any concerns that can remain even in this seemingly straightforward application?
VG: The only thing here which is important is that while the final evaluation is some composite metric, there should be some voting (how to measure/set the right weights or impact of different metric parts? Like materials quality vs price vs time to complete). But, even taking into account such questions, it is a very important first step, which definitely allows improvement of the quality and prevention of human bias, especially in procurement.
What are the broader implications if AI agents become widespread in government functions?
VG: It is important to consider the “what if” perspective. So, what if, let’s say, 50% of countries will involve AI agents even up to the replacement of some realistically possible functions/ministries, etc. This is quite a huge area for any imagination or even fiction books, starting from agents-to-agents communication on the country level (and who will be more efficient in agent tuning/instructions) and ending with the error values/severity and fluctuations in citizen trust, which can create cycles (higher trust – more agents, errors and lower trust – fewer agents).
Related insights
