The second edition of “alGOVrithms. The State of Play” is taking an in-depth look at the usage of Automated Decision Making in government-citizens relations in Czechia, Hungary, Kosovo, North Macedonia, and Poland. Research done were examining the existence of relevant legal frameworks, the practice of creating, implementing, and using of Automated Decisiob Making processes on different levels of government, focusing on their transparency, accountability, and potential risks of discrimination. During this research, the authors of the report were also particularly interested in understanding how alGOVrithms (the term created by the research team to describe ADM systems used by central and local authorities) are being procured and who is involved in the process of their implementation.
The examples of alGOVrithms presented in the report are sorted by the purpose of their usage. We have identified the tools which support the process of allocating judges and other public officials to specific cases; to detect frauds and misuse of public funds; facilitating the administrative and procurement processes and granting allowances and other social benefits. We have also detected the examples of alGOVrithms especially designed to respond to the challenges connected with the COVID-19 pandemic. Among others, the reader will find descriptions of tools supporting authorities in controlling the observance of the quarantine, accuracy of spending of state aid by selected companies, or detecting risks of spreading contamination.
As in 2019, we have not identified the existence of comprehensive legal frameworks regulating the process of alGOVrithm implementation and usage and their transparency. Apart from modest examples, we have also not encountered the practice of engaging CSOs’, academia, or other independent experts in the process of elaborating policies and regulations, as well as working on the specific technological solutions. As in previous years we found it difficult to gather the necessary knowledge on alGOVrithms due to the general lack of transparency which is particularly visible in the case of applying IT solutions in the public sector. We have also identified the significant problem of the lack of competencies of public officials who are directly responsible for using alGOVrithms. Very often they have to rely on the expertise of external companies who are delivering tools and there are no comprehensive internal processes that mitigate risks of negative impacts generated by Automated Decision-Making processes.
It does not mean that authorities are completely indifferent to these challenges. Some countries have prepared in recent months quite complex strategies on Artificial Intelligence development in which they are signaling the need of preparing regulatory frameworks, ethical guidelines, or concrete mechanisms as Algorithmic Impact Assessments. In these documents, we can also find the promise of creating multi-stakeholders’ bodies that will oversight the process of Automated Decision-Making implementation and other mechanisms that will increase the transparency and accountability of alGOVrithms. Still, most of the concrete work is ahead of governments in countries that were the subject of the research, and there is little reflection among the public officials on the need to create similar strategies around the simpler examples of algorithms, which are not using AI but still have a significant impact on rights and obligations of citizens.
For these reasons, we are urging governments to follow our policy recommendations which are the following:
- Establish a regulatory government body that overlooks the development, implementation, and usage of alGOVrithms;
- Train civil servants with the necessary skills to monitor and audit alGOVrithms;
- Develop a legal framework to guide the development of trustworthy alGOVrithms and strengthen their transparency;
- Create the necessary legal framework and incentivize CSOs and academia to engage in monitoring and regulating alGOVrithms;
- Initiate public debates between policy-makers, CSOs, media, businesses, and citizens to co-create lawful alGOVrithms that guarantee good governance, human rights, and democracy principle;
- Create a national, high-level expert group on alGOVrithm and AI, comprised of legislative, CSO, business, and academia representatives, which will act as counsel to government bodies attempting to utilize ADMs;
- Incentivize the share of knowledge and exchange of good practices between government agencies on the responsible use of alGOVrithms;
- Develop national standards for procuring, supervising, and auditing alGOVrithms