National Audit Office: artificial intelligence could save up to 30 percent of working time

• Applying artificial intelligence (AI) could save up to 30 percent of working time for the state, but only 15 percent of institutions use it.
• There is a lack of infrastructure in the Lithuanian public sector to support AI systems.
• Public sector bodies do not have clear data management processes, do not adhere to standards for AI management, and staff lack AI competences and training.
• In the period 2021-2030 EUR 42.35 million is earmarked for the development of Lithuanian language datasets that enable AI systems to understand and communicate in Lithuanian. However, the public sector still does not have all the resources needed to develop the AI, and innovation in this sector is stagnating.

Artificial intelligence is already making a difference in business, science and everyday life. But the public sector is still stagnating. The National Audit Office's audit Management of artificial intelligence in the public sector shows that the biggest obstacles to a breakthrough in this area are unresolved issues of infrastructure, data and staff competences.

"Artificial intelligence can and must become an important tool in the Lithuanian public sector. It would help us work faster, more efficiently and more cost-effectively. Reducing the administrative burden, improving customer service, making public procurement more efficient, and more accurate forecasts are just a few examples. A well-coordinated, systematic deployment of AI can save hundreds of millions of working hours and, consequently, generate millions of euros in economic benefits for our country. But without a clear strategy and proper preparation, these possibilities remain just visions", notes Auditor General Irena Segalovičienė.

The audit results show that there is a lack of infrastructure in Lithuania to support the operation of AI systems: needs are not assessed and monitored, and there is no national plan for the development of AI capacities. Public sector bodies do not have clear data management processes and do not comply with standards for AI management. There is a lack of AI competences and training for staff in this sector. In order to fully exploit the potential of AI in the public sector, it is necessary to ensure that it is able to recognise and work with the Lithuanian language. This requires linguistic resources such as text and linguistic datasets, dictionaries and models that enable AI to understand and process natural language.  High-quality language resources can improve the efficiency of services, reduce manual work and the risk of human error, while their lack hinders the adoption of innovative AI solutions in the public sector.

In the period 2021-2030 EUR 42.35 million is earmarked for the development of language resources for public sector AI solutions. However, the public sector still does not have all the linguistic resources it needs to develop the AI and innovation in the sector is stagnating. Although most of the foreseen funding has already been allocated, there are no tangible results yet. The public sector is not expected to be able to use the necessary language resources until 2026.

What does the National Audit Office recommend?

In order to ensure that the public sector keeps pace with technological progress, is able to deploy AI solutions more efficiently and ensures data security, the National Audit Office made recommendations to the Ministry of Economy and Innovation. The Ministry formulates and coordinates the implementation of the policy on information resource management and digital development. The recommendations of the National Audit Office underline the need to:

• strengthen monitoring of computing capacity;
• ensure the quality of data management;
• provide training and methodological support to public sector staff;
• ensure the availability of high quality and comprehensive language resources.

These steps would remove key barriers and accelerate the deployment of AI in all areas of the public sector, from health and education to social security and public administration.

The implementation of the National Audit Office's recommendations would not only provide a strategic approach to the application of AI technologies, but also improve the legal framework in this area and create the conditions for effective management of AI technologies in the public sector. Furthermore, proper management of the resources required for AI functionality will ensure the efficient implementation of projects related to the deployment of AI solutions in the public sector.