At this time, Lithuania does not have an Artificial Intelligence (AI) transparency standard that would help oversee the algorithmic tools used by public sector institutions and understand their purposes. In general, the country lacks a strategic vision, clear commitments, and accountable institutions that seek to mitigate the risks of AI technology while promoting its benefits, according to an exploratory transparency study conducted by the “Transparency International” Lithuania branch (TILS).
Public sector institutions in Lithuania currently use algorithmic tools to augment the efficiency of internal processes and improve monitoring and evaluation. Every 4th institution does this (12 of 53). Fully automated decision-making, where final decisions are made without human input, is not yet in use.
According to the study participants, some of the main risks associated with the use of algorithmic tools are algorithmic bias, insufficient testing, incorrect use of data sets, and failure to account for country contexts. The most notable benefit of the use of these tools is the reduction of mechanical, routine work.
According to surveys, 4 of 12 institutions (Ministry of National Defense, National Courts Administration, State Tax Inspectorate, and the Public Procurement Office) conduct continuous monitoring to ensure that decisions made by algorithms are accurate and ethical.
Public sector employees often lack knowledge about the use and possibilities of artificial intelligence, which may pose a risk to the quality of decisions made. For example, there is insufficient investment in helping employees understand the capabilities of AI impact assessments and monitoring, or how to properly prepare technical specifications for these tools when purchasing them from suppliers.
“Artificial Intelligence is an inevitable part of decision-making in the public sector. That is why I would especially like our government representatives to take initiative and clearly agree on the strategy and transparency standard for the use of artificial intelligence in Lithuania. We have to know what lies behind the decisions that directly affect our lives, and how they are reasonably accepted,” said Ingrida Kalinauskienė, head of the “Transparency International” Lithuania branch.
“The Government of the Kingdom of the Netherlands was one of the first in Europe to start using algorithms in the decision-making process. However, we have learned the hard lesson that technology can bring speed but not necessarily accuracy. People should have the opportunity to make final decisions on public management questions, not algorithms,” proposes H.E. Ambassador of the Kingdom of the Netherlands to Lithuania Jack Twiss Quarles van Ufford.
Considering the main insights of the research and the visible risks, TILS recommends:
- The Ministry of Economy and Innovation should become actively involved in and promote the conversation about the strategic creation and use of artificial intelligence on the national scale. In addition, gather the professional community to discuss and agree on a national strategy for the use of AI, the transfer of EU AI regulation, the creation of AI transparency standards, and the foundation or assignment of a responsible institution in Lithuania.
- Regularly collect and publish summarized statistics about the use of AI technology in Lithuania’s public sector that show which AI technologies are used, for what purposes, and how much funding is allocated for them.
- Create an AI transparency standard that would ensure that data about the use of AI in public sector institutions is publicly available and easily accessible in an open data format.
- Raise the competencies and knowledge level of employees of public sector institutions looking to formulate and/or utilize accurate AI decisions, in order to properly use and evaluate their effect and created change.
- Public sector institutions using automated decision-making should regularly assess whether they are operating accurately and ethically. Establish feedback mechanisms where necessary.
This study and recommendations are based on 7 semi-structured in-depth qualitative interviews with representatives of public sector institutions and academia, conducted by TILS from September-October 2023, and official requests (55) sent in June 2023 to obtain information by data.
You can find a summary of the results here and can learn more about the study here.
More information: Ingrida Kalinauskienė, info@transparency.lt, +370 5 212 69 51
The initiative is implemented in cooperation with the Embassy of the Kingdom of the Netherlands in Lithuania.