Different Levels of Access to Artificial Intelligence (AI) – A Deep Dive into OpenAI and Deepseek

- Bondarenko, Tetiana (orcid.org/0000-0001-9879-0319), Koeberlein-Kerler, Juergen (orcid.org/0000-0002-9205-5951), Shteimiller, Iryna (orcid.org/0009-0007-9488-8221), Rostoka, Marina Lvovna (orcid.org/0000-0002-1891-5482) and Fedorova, Yuliia (orcid.org/0000-0002-9381-1229) (2026) Different Levels of Access to Artificial Intelligence (AI) – A Deep Dive into OpenAI and Deepseek Lecture Notes in Networks and Systems (1846). pp. 248-255. ISSN 2367-3389

Text Rostoka ML_article_2_2026.pdf Download (178kB)

Abstract

The rapid development of artificial intelligence (AI) technologies is transforming all spheres of life - from education and healthcare to industry and public administration. AI systems, in particular generative models (such as GPT, Deepseek, Claude, Gemini), are increasingly being used not only by researchers and large companies, but also by the general public. However, there is a significant difference in the level of access to these technologies: from open source models to closed solutions with limited functionality, which creates digital inequality and limits the innovation potential of certain communities. Against this background, it is of particular interest to compare the OpenAI platforms and Deepseek as an example of an open, research-oriented initiative. Studying their functional, ethical, and accessibility differences allows us to better understand the architecture of the modern artificial intelligence landscape. The study used a comparative and systematic analysis of two platforms: OpenAI, as an example of a commercial but widely available AI system with a high level of integration into mass-market products (ChatGPT, Copilot, API); Deepseek, as a representative of an open research initiative focused on shared knowledge and transparency (Deepseek-Coder, Deepseek-VL), which allows downloading models, playing them locally, and adapting them to one's own needs. AI platform access criteria were developed and the features of each criterion were described: Architectural features of models (e.g., Transformers, multimodality); License terms (opensource vs. closed-access); Accessibility for educational and research purposes; Level of documentation, support, community activity; Ethical aspects (transparency of algorithms, data protection, control over use).

Item Type:	Article
Keywords:	Access Levels, Artificial Intelligence, Open Models, Systems Analysis.
Subjects:	Science and knowledge. Organization. Computer science. Information. Documentation. Librarianship. Institutions. Publications > 00 Prolegomena. Fundamentals of knowledge and culture. Propaedeutics > 001 Science and knowledge in general. Organization of intellectual work > 001.8 Methodology Science and knowledge. Organization. Computer science. Information. Documentation. Librarianship. Institutions. Publications > 00 Prolegomena. Fundamentals of knowledge and culture. Propaedeutics > 004 Computer science and technology. Computing. Data processing Science and knowledge. Organization. Computer science. Information. Documentation. Librarianship. Institutions. Publications > 3 Social Sciences > 37 Education > 378 Higher education. Universities. Academic study
Divisions:	V. Sukhomlynskyi State Scientific and Pedagogical Library of Ukraine > Department of Scientific Information and Analytical Support of Education
Depositing User:	завідувач Марина Львівна Ростока
Date Deposited:	03 Jun 2026 07:15
Last Modified:	03 Jun 2026 12:29
URI:	https://lib.iitta.gov.ua/id/eprint/749114

Downloads

Actions (login required)

View Item