The Moje Państwo Foundation discussion panel entitled "AI / ADM systems in the Central and Eastern European (CEE) public sector: what can (or has) gone wrong in terms of citizens' rights?" was held on June 8, 2022, during the international RightsCon 2022 conference organized by Access Now.
The discussion panel was attended by:
Monika Kajalidis, COO at Moje Państwo Foundation;
Sandor Lederer, co-founder and director of K-Monitor;
Holger Zscheyge, president of the European Legal Technology Association (ELTA), founder of Infotropic Media;
Magdalena Siwanowicz-Suska, legal director at Moje Państwo Foundation (as moderator).
During the panel, various aspects of the development of AI/ADM systems in the public sector were discussed, based on the examples of Poland, Hungary and Russia.
Issues raised included:
1. present situation regarding the implementation of national AI strategies in the region of Central and Eastern Europe (CEE);
2. political and economic aspects of AI implementation in the CEE region (including the impact of the war in Ukraine on the development of AI in Russia);
3. good practices and examples of AI / ADM implementations in public administration and controversies surrounding such systems;
4. risks for citizens' rights related to AI in the public sector.
- Sandor Lederer presented the situation of implementation of AI/ADM systems in Hungary. Hungarian AI strategy was published in 2020. The document is more technological in nature and largely concerns the development of artificial intelligence in the private sector. In Hungary, awareness of AI among officials is average and the level of use of technology is very low. The collecting of personal data in the digital market can be described as being monopolized.
- Monika Kajalidis highlighted the proactive approach of the Polish government administration to the implementation of AI policy and the visibility of public sector activities to the public. In Poland, for example, there is a dedicated website where you can find a database of implemented public innovations together with a map of innovations in the regions of Poland. A special Working Group on Artificial Intelligence has been set up at the Chancellery of the Prime Minister. The Moje Państwo Foundation participates in its work. The strategy for the development of AI in Poland describes in detail the steps of the development of artificial intelligence, as well as indicates the short-term, medium-term and long-term strategy objectives. The document is divided into 6 thematic areas and provides guidance for businesses and the public sector.
- Holger Zscheyge said that the strategy for AI development in Russia was published in 2017. The document outlines ambitious plans and goals for the implementation of artificial intelligence technology by 2030. Russia aspires in the document to become a global leader in the field of artificial intelligence. The highest concentration of enterprises using AI technologies are in Moscow (79%) and St. Petersburg (10%). There is a big gap in the development of AI technology in other regions of Russia. The private sector does not collaborate much with the public sector in the development of artificial intelligence technologies. The war in Ukraine will significantly affect the current situation of the innovation market in Russia because of sanctions, the relocation of large companies - such as Yandex, the outflow of qualified IT personnel and restriction on access to technology and equipment. According to Holger Zscheyge, we will thus be able to observe a reduction in investment in this sector of the economy.
- In Hungary, a good example of the use of AI in the public sector is the system used by the tax administration to prevent fraud and money laundering. A controversial example of the use of AI algorithms was the use of a questionnaire in schools for children to plan their future careers. Another example that needs discussion is the use of AI in CCTV (Closed Circuit TeleVision) systems in Hungary for street monitoring and facial recognition.
- In Russia, there is no information about controversial use of AI algorithms in the public sector. Most projects are terminated due to the use of a poorly selected dataset or due to a lack of sufficient investment in this sector. AI algorithms are commonly used in City Contact Centers (due to the lack of adequate human resources). The AI algorithm can answer 50 thematic issues - for example, regarding obtaining financial support or on the passport exchange procedure. The hotline receives up to 65 million calls a day. Artificial intelligence in Russia in the public sector is also used to control public transport drivers and to monitor streets and detect crimes.
- In Poland, the Central Anti- Corruption Bureau has implemented a system to control the correctness of granting subsidies to enterprises in connection with Covid19. Finding the information about the system was a big challenge. The Ministry of Justice uses algorithms to assign cases to judges in ordinary courts. The Foundation's inquiries regarding access to public information on how the random case allocation system works did not have any impact at the first stage. The case was referred to the provincial administrative court, but only further proceedings and a favorable verdict of the Supreme Administrative Court (NSA) for the Foundation led to the information about the system being made available. The decision of the Supreme Administrative Court that information about the algorithm of the system used by public authorities is public information is a great success of the Foundation. In the future, the ruling may be a significant point of reference in terms of access to information on AI/ADM algorithms used in public administration. All documentation on the system of random allocation of cases used in the judiciary has been published on the website of the Ministry of Justice.
- Panel participants unanimously concluded that information on artificial intelligence and implemented AI/ADM algorithms should be common, transparent and accessible. A harmonized system of using AI systems in public administration should be created, in particular in the process of public procurement. An important aspect is also the creation of unified webpages with databases where the citizen can find information about the applied public AI systems and algorithms. An important issue is also increasing the knowledge of AI among civil servants and specialists in public administration and the method of overseeing the AI algorithms. The issue of legal ownership of algorithms in public companies remains unresolved.
- Foundation Moje Państwo published a report on the use of AI/ADM in public procurement - guidelines for the public sector. The document was met with great interest and was published on the website of the Ministry of Digital Affairs.
- According to the participants, work on AI ACT in the European Commission requires the involvement of experts and civil society in the discussion on amendments, introducing transparent, trustworthy and understandable principles of using AI/ADM systems for the society by public administration.
Written by Iwona Karkliniewska, Researcher AI at Foundation Moje Państwo