No authority can be an “omnipotent” – that is, have deep expertise in journalism, finance, electoral law, industry, and information technology at the same time. Such an approach would lead to excessive centralization of regulatory power and disproportionate intervention in various social areas.
Alban ZENELI
The regulation of so-called Artificial Intelligence (AI) in Kosovo should be built on the principle of strengthening existing sectoral regulators, providing them with a clear legal basis and specific competences to address the use of this technology within their scope. This approach is more sustainable and proportionate than the creation of a single central authority for regulating AI, which would attempt to establish uniform rules for all sectors, without taking into account the fundamental differences between them.
The creation of a single authority to regulate AI would pose a real risk of overregulation and infringement of fundamental freedoms, especially in sensitive sectors such as media, finance and electoral processes. No authority can be an “omnibus”, that is, have deep expertise in journalism, finance, electoral law, industry and information technology at the same time. Such an approach would lead to excessive centralization of regulatory power and disproportionate intervention in different areas of society.
For example, in the media sector, the relevant regulator is the Independent Media Commission (IMC). If AI regulation were entrusted to a single central authority, in the absence of sectoral expertise, there is a risk that the use of AI in the production and distribution of media content would be treated in a repressive manner and would undermine freedom of expression, media pluralism and editorial independence. Instead, the IMC should be provided with a legal basis to regulate the use of AI in the media by: requiring transparency of AI-generated content; setting standards for clear labelling of AI-generated media content; preventing AI-generated disinformation and protecting the rights of the audience/public in relation to the media. This would appropriately replicate the self-regulation already undertaken by the Kosovo Council for Online Media for this technology.
Similar to the media sector, in the area of personal data protection, the Information and Privacy Agency (AIP) is the key institution for the protection of personal data. Given that AI relies on massive data processing, this institution should be strengthened with powers to oversee automated decision-making on this data, prevent discrimination by data-based algorithms, and monitor the compliance of AI systems with data protection standards. Perhaps adding to this the example of countries that have included photographs, voices, and other data of citizens within the scope of copyright, protecting them from the possibility of reprocessing with AI technology.
As part of this sectoral approach, the Central Election Commission (CEC) should be given special legal authority to oversee and sanction the use of Artificial Intelligence in political campaigns and electoral processes. AI is increasingly being used to create political content, target voters, and automate electoral communication, which can directly impact the integrity of elections.
Also, in the context of the fight against politically based disinformation, especially in election campaigns, the Central Election Commission should have the power to sanction political parties that do not use AI in a transparent manner, especially in cases where political content generated or manipulated by AI is not clearly declared as such, or algorithms are used for political targeting without basic public information.
This sectoral approach to regulating this technology would help all institutions to set clear standards for the transparent, responsible and ethical use of AI. On these principles, regulation should continue in all regulatory institutions such as the Regulatory Authority for Electronic and Postal Communications (ARKEP), the Central Bank, educational and research institutions, universities and the Academy of Sciences. The sectoral solution to regulating AI offers a much more professional approach in the relevant fields, more democratic and essentially more functional than the approach of regulation by an omniscient authority whose potential knowledge would be truncated in each field.
Excerpt from the roundtable discussion "The Role of Artificial Intelligence in the Field of Information Integrity", organized by ASD/Hybrid.info.
(Alban Zeneli is a lecturer at the Department of Journalism, Faculty of Philology at the University of Pristina. He is also the author of many scientific publications in prestigious international journals. Zeneli is a regular columnist at hibrid.info).