CIS Countries Need Common Approach to AI in Science and Education
The Interparliamentary Assembly of the CIS held a round table on 22 October at ExpoForum in St. Petersburg to discuss legal regulation of artificial intelligence in science and education. Parliamentarians and experts compared approaches across CIS countries and considered principles for possible model legislation, including transparency, proportionate regulation and shared responsibility. Participants noted that no CIS country yet has a comprehensive AI framework tailored to science and education, and proposed safeguards for research verifiability, data and intellectual property protection, and testing new technologies. Proposals from the discussion will be submitted to the IPA CIS commissions and expert advisory board for further consideration.
A round table organised by the Interparliamentary Assembly of the CIS examined legal regulation of artificial intelligence in science and education. Participants compared national approaches and discussed principles for possible model legislation.
On 22 October, a round table discussion on the legal regulation of artificial intelligence in science and education across the CIS was held at ExpoForum in St. Petersburg. The event was part of the St. Petersburg International Scientific and Educational Salon, which focuses on vocational education, science and innovation in the 21st century.
The event was organised by the Interparliamentary Assembly of the CIS and brought together parliamentarians, academics, legal professionals and experts from across the Commonwealth.
Discussions focused on a number of pressing questions. How can AI be used to develop educational materials and conduct research? Where is the line between legitimate AI assistance and academic misconduct? Who owns the rights to AI-assisted work? And who is liable for decisions made with the help of algorithms?
Secretary General of the IPA CIS Council Dmitriy Kobitskiy noted that the Assembly had extensive experience in developing model legislation in related areas, including science, education, innovation and data protection. This provides a solid foundation for identifying the additional legal safeguards needed in the scientific and educational sectors, he added.
The discussion was chaired by Chair of the IPA CIS Permanent Commission on Science and Education and Chair of the Committee of the Federation Council of the Federal Assembly of the Russian Federation on Science, Education and Culture Lilia Gumerova and Chair of the IPA CIS Expert Advisory Board on Science and Education and Rector of Peter the Great St. Petersburg Polytechnic University Andrey Rudskoy.
Lilia Gumerova outlined the key provisions of the IPA CIS Model Law on Artificial Intelligence Technologies. The law establishes a framework for regulating relations arising at every stage of the AI lifecycle, with particular emphasis on ethical standards in the development and use of AI.
She also highlighted the need to explore how the Model Law could be adapted to academic settings, identify gaps in the existing framework and develop proposals for future IPA CIS model legislation.
Experts at the discussion noted that advances in generative AI are presenting new challenges for both education and the legal system, calling into question established legal assumptions and traditional approaches to ensuring academic integrity.
One of the main challenges is determining what constitutes independent work. The use of AI does not, in itself, mean that a student or researcher has failed to work independently. Neural networks can serve as valuable research assistants, for example, by helping users find and analyse scientific literature.
When considering how universities should regulate the use of AI, participants highlighted several guiding principles:
- Technological neutrality. The independence and reliability of academic work should be assessed on the basis of how AI is used, rather than simply whether it has been used.
- Proportionate regulation. Blanket bans are largely ineffective in practice, while requiring specific AI systems to be labelled in all cases may not be justified.
- Transparency. Institutions should set out clear expectations for disclosing AI use in their ethics policies and self-regulatory frameworks.
- Shared responsibility. Standards for the responsible use of AI should apply to academic staff as well as students.
Participants also examined the concept of “layered co-authorship”, which recognises the different parties involved in generating AI-assisted content: the owners of the original works used to train a model, the developers of the AI system and the end user or prompt engineer. The user’s creative input and the specificity of their prompts can significantly shape the final output.
This form of interaction was likened to “automated folklore” and discussed in the context of emerging risks, including the coordinated use of AI agents to conduct cyberattacks or commit fraud. These developments raise difficult questions about how legal responsibility should be allocated among the parties involved.
Deputy Chair of the IPA CIS Expert Advisory Board on Science and Education Dmitriy Mokhorov then outlined the current state of AI regulation across the CIS.
Azerbaijan is implementing its AI Development Strategy for 2025–2028, which provides for requirements for responsible AI systems and the introduction of a regulatory sandbox.
Armenia combines its international commitments with targeted regulation and has adopted legislation addressing synthetic content and deepfakes.
Belarus is drafting a dedicated national law on AI, alongside a broader regulatory framework.
Kazakhstan has adopted a dedicated AI law introducing risk management, auditing, labelling requirements for AI-generated content and liability for harm. AI tools are also being introduced into court proceedings.
Kyrgyzstan is developing an ethical framework through a draft regulation on AI ethics. The proposed rules explicitly state that AI has no legal personality and that responsibility rests with humans.
Russia has adopted framework legislation focused on large foundational models, with more detailed rules to be established through secondary legislation. Courts have also applied existing procedural rules to cases involving the use of AI-generated content that has not been properly verified.
Tajikistan is pursuing an AI development strategy through 2040, with a particular focus on integrating AI into education.
Turkmenistan is developing rules, standards and technical regulations governing AI.
Uzbekistan has opted for targeted legislative amendments, including a prohibition on making decisions affecting human rights solely on the basis of AI-generated conclusions. The country has also approved AI ethics rules and is preparing a separate draft law.
“Although their approaches differ, all CIS countries are moving towards establishing formal rules for AI. However, none yet has a comprehensive regulatory framework specifically tailored to science and education. Bridging this gap is precisely where IPA CIS model legislation can make a difference,” Dmitriy Mokhorov stressed.
Any future framework, participants noted, should be based on several key principles: respect for the digital sovereignty of CIS countries, recognition of the specific needs of science and education, and the principle that only human beings can be subjects of legal relations and bear professional responsibility.
A dedicated model legislative act on the use of AI in science and education could establish safeguards for the verifiability of research, the protection of data and intellectual property, and the flexible testing of new technologies.
Lilia Gumerova said that the proposals and issues raised during the discussion would be submitted to the IPA CIS Permanent Commission on Science and Education and the IPA CIS Expert Advisory Board on Science and Education for further consideration.
Deputy Secretary General of the IPA CIS Council – Plenipotentiary Representative of the Kuryltai of the Republic of Kazakhstan Ashat Mukashev also took part in the discussion.
https://iacis.ru/News/IPA_CIS_Commissions/CIS_Countries_Need_Common_Approach_to_AI_in_Science_and_Education