Artificial intelligence is rapidly changing how scientific knowledge is produced, evaluated, shared, and accessed. AI systems support researchers in searching literature, analyzing data, generating hypotheses, and communicating findings, providing new opportunities for making research knowledge more accessible and understandable. At the same time, AI use in science and science communication raises fundamental questions about bias, reliability, transparency, responsibility, and equitable access to knowledge. This theme explores these opportunities and threats, conceptualizing AI tools as interfaces at the nexus of scientific practices and access to scientific knowledge. It builds on the human right to science and the living guidelines for responsible generative AI use, as well as similar normative documents developed by the EU, OECD, and UNESCO.  

This theme addresses the following challenges: 

1. Access to scientific knowledge: AI may lower barriers to information, but it can also reproduce inequalities, obscure sources and exclude people whose languages, knowledge practices or needs are insufficiently represented. A human-rights-based approach to interfaces based on the human right to science will be developed, including elements such as openness, participation, accessibility, and accountability. 

2. Science fact checking: AI-generated outputs can appear convincing even when they are inaccurate, unsupported, or misleading. This theme explores how researchers, professionals, and wider publics can assess AI-supported scientific claims, for example, through science fact-checking games and other participatory interfaces that make critical evaluation accessible and engaging.