TY - GEN A1 - Koperwas, Jakub A1 - Rybiński, Henryk A1 - Szewczyk-Kłos, Danuta T1 - From Network of Research Information Systems to AI-powered Accessible Repository - Open Science Infrastructure in Poland T2 - Themenkreis 5: Digitalität und KI/ Wie offen sind wir? (24.06.2025, 13:00 - 15:30, Halle 4.1./ Raum V) N2 - Presentation focuses on building open research infrastructure by universities in Poland. The process started with Research Information System, developed at the Warsaw University of Technology . The main idea was to promote transparency in research which involves publishing real and up to date information about all research-related activities. This resulted in a knowledge base that provides the functionalities of institutional repository of research data, publications, patents, etc., supporting the university authorities in strategic decision support processes, internal reporting including researchers' assessment, reporting to the National Authorities and European bodies such as OpenAIRE. Currently there is a network of institutional CRIS systems (currently 50+), as the bottom-up initiative of the Polish university libraries. The steering group established under the umbrella of the Conference of Directors of Polish Academic Libraries takes care of the system development and data standardization by curating common databases.The direct follow up was a Polish Platform for Medical Research, where the medical Universities decided to promote their medical-related achievements in a single platform including fully accessible full-text files. This unrevealed the problem of accessibility of PDF files for individuals with visual or other disabilities. PDFs can contain a wide range of content, from simple text to complex multimedia elements. Providing alternative text for images, ensuring a proper document structure, and implementing meaningful link text require careful attention to detail. The complexity increases further with the inclusion of tables, forms, and multimedia elements.Therefore another research project conducted and co-financed by the EU was rised to solve the problem with the help of AI. The solution that was built, offers a way to address the challenges associated with manual annotation by providing efficiency, accuracy, and adaptability. Y1 - 2025 ER -