Dejan Mircetic is a research scientist at the Institute of Artificial Intelligence of Serbia, where he currently works on projects related to the application of AI solutions to business and industry problems. Established in March 2021 by the Serbian government, the Institute for Artificial Intelligence of Serbia brings together AI enthusiasts – scientists, researchers and industry experts – dedicated to AI research.
In an interview with The Recursive, Mircetic discusses how commercialization can fuel technological progress in AI, the intersection between scientific research and practical implementation, the application of AI in various fields and more Again.
The following interview was conducted as part of The Recursive’s “State of AI in CEE” report. Download the full report with insights from over 40 experts and analysis of 900 CEE AI product companies here.
The Recursive: How do you see the intersection between scientific research and industry commercialization today?
Dejan Mircetic: Industry commercialization not only provides the funds and resources needed to pursue cutting-edge research, but also serves as a feedback loop, highlighting real-world challenges that require innovative solutions. This synergy propels both scientific discovery and entrepreneurial ventures.
However, it is crucial that as we move forward, the integrity of the research remains intact and that commercial benefits are balanced with ethical considerations. Essentially, the blend of pure science and its commercial potential is what will shape the future of AI and the broader technology industry in Central and Eastern Europe.
What are the current challenges or problems in CEE sectors that AI can address or solve?
One of the main challenges is the modernization of existing systems, especially in the manufacturing and utility sectors. AI can help streamline operations, optimize supply chains and predict maintenance needs. Another pressing concern is the efficient use of resources in agriculture. Advanced AI models can contribute to precision agriculture, enabling farmers to use water, fertilizers and pesticides more wisely, improving yield and sustainability.
What are the potential risks and challenges associated with the adoption of AI in the industry?
The adoption of AI in industry, while offering enormous benefits, comes with several risks and challenges. One of the most pressing concerns concerns ethical considerations. When AI systems make decisions, there is a risk of bias, especially if the data they are trained on is unrepresentative or contains inherent biases.
This can lead to unfair or discriminatory outcomes in sectors like finance, healthcare or recruitment. Another challenge is the potential loss of jobs. AI-enabled automation could lead to job losses in some sectors, requiring a focus on workforce upskilling and reskilling.
Then there is the issue of data privacy. Since AI systems often rely on large amounts of data, it becomes paramount to ensure that personal and sensitive data is protected and not misused.
What are your personal concerns regarding the ethical implications of AI in your particular area of research?
In my work in time series forecasting and supply chain analysis, the ethical implications of AI weigh heavily on my mind. A key concern is the potential misuse of predictive data, which could lead to biased decision-making, particularly if the underlying data is unrepresentative or has inherent biases.
Additionally, as supply chains inherently involve multiple stakeholders, transparency of AI-based decisions becomes paramount. Without clear communication about how certain predictions are made, there could be a lack of trust between partners, or even an unintentional preference of one entity over another.
How would you assess the availability and quality of talent with the necessary AI skills in Serbia and the region?
Assessing the AI talent pool in Serbia and the wider CEE region, it becomes clear that there is a rich reservoir of qualified individuals. Historically, this region has strong foundations in mathematics, engineering and computer science, providing an advantageous starting point for developing AI expertise. Higher education institutions in Serbia have gradually introduced and updated study programs in AI and data science, demonstrating the growing talent base.
However, while there is an abundance of raw talent, there remains a deficit in terms of specialist skills and practical industry experience. In the field of AI, companies and research institutions are seeking expertise in areas such as time series forecasting, deep learning, natural language processing, and computer vision. But beyond technical skills, there is a growing demand for professionals who can understand the ethical and societal implications of AI, highlighting the need for interdisciplinary training.
What are the main specializations and skills in AI in Serbia according to your findings so far?
In my research on time series forecasting and supply chain analysis in Serbia, several AI specializations and skills emerged as important. Above all, expertise in deep learning, particularly recurrent neural networks (RNN) and long-term and short-term memory (LSTM) networks, is highly sought after due to their proficiency in processing sequential data such than time series.
The interpretability of machine learning models is another crucial area. Since supply chains can be complex and involve multiple stakeholders, understanding and explaining AI-based decisions is critical to fostering trust and collaboration. Additionally, optimization techniques, especially those related to nonlinear programming and constraint optimization, are crucial as they directly impact supply chain efficiency. This is closely related to operations research skills, which are an integral part of many supply chain issues.