Exploring the Potential of AI in Government Decision-Making

a photo of the U.S. Capitol Building using a long exposure lens with streaming colors of cars and traffic lights.

By Michael Parker

ALBANY, N.Y. (Dec. 1, 2022) — Artificial intelligence (AI) has become an important tool for governments around the world. However, it isn’t clear the extent to which AI can improve decision-making, including within policy domains and government functions such as the public budgeting process.

Now, researchers at UAlbany’s Center for Technology in Government (CTG UAlbany) and Universidad Autónoma del Estado de México have conducted a study aimed at answering how AI techniques can be used to boost government decision-making by gearing public spending toward increasing GDP, decreasing inflation and reducing income inequality.

Publishing in a recent issue of Government Information Quarterly, CTG UAlbany Director J. Ramon Gil-Garcia and co-authors David Valle-Cruz and Vanessa Fernandez-Cortes have authored the article, “From E-budgeting to smart budgeting: Exploring the potential of artificial intelligence in government decision-making for resource allocation.”

To complete their analysis, the authors used a multilayer computer model and a multi-objective genetic algorithm to analyze World Bank Open Data from 1960 to 2019, including 217 countries.

Portrait of J. Ramon Gil-Garcia, Director of CTG UAlbany.
Center for Technology in Government Director J. Ramon Gil-Garcia

“The advantages of implementing this type of decision support system in public expenditures allocation arise from the ability to process large amounts of data and to find patterns that are not easy to detect, which include multiple non-linear relationships,” said Gil-Garcia, who serves as a professor of Public Administration & Policy at UAlbany’s Rockefeller College of Public Affairs and Policy. “Some technical aspects of the expenditure allocation process could be improved with the help of these kinds of techniques. In addition, the results of the AI-based approach are consistent with the findings of the scientific literature on public budgets, using traditional statistical techniques.”

Indeed, AI has become an essential issue on the agenda of governments around the world due to its potential benefits and positive implications for efficiency, transparency, service quality, and public value. However, the AI black box and the lack of explainability of some deep learning techniques could result in lack of trust, inequity, bias, the massive replacement of the workforce (particularly in routine tasks), and the increase in the digital divide.

Despite the potential for positive or negative impacts in the use of AI in government, the authors believe it is clear that AI techniques could assist in decision-making by supporting public managers and government officials with simulations, new ideas, and innovative approaches to better understand data and the dynamics among multiple variables.

Technology use in government (digital government) has enabled public services delivery through the Internet, fostered efficient capture, process, and report on data, and improved decision-making. However, advances in smart technologies, better informed and connected citizens and globally connected economies have created additional opportunities.

The authors also found that governments have begun to take the concept of digital government to a new level as they realize the power of data and heuristic processing through artificial intelligence to improve their services, interact with citizens, develop policies and implement solutions for the welfare of the community and become a smart government.

Smart government is considered one of the key trends that governments are following with the participation of the public and private sectors, as well as NGOs and members of civil society, boosted by technology, like the Internet, Big Data, open data, IoT, and AI, bringing the potential to improve planning and decision-making in governments.

“Comparing the AI results with what has already been done on budgeting, we learned that the most important aspects that can generate an efficient public budget refer to policies that improve GDP, income inequality, and inflation reduction,” said Gil-Garcia. “These aspects are based on strategies focusing on improving public debt, fostering education and economic saving, investing in the agricultural sector, and addressing the population growth and public health problems. According to our AI-based analysis approach, the three most essential budget expenses that could improve the GDP are public spending, agriculture, and subsidies.”

The findings were consistent with previous research that indicates that agriculture and public spending have a significant and positive effect on the GDP, and health directly affects productivity.

Gil-Garcia’s co-authors for the study were David Valle-Cruz and Vanessa Fernandez-Cortez, both of Universidad Autónoma del Estado de México.