CC BY 4.0Chugunova, MarinaMarinaChugunovaHarhoff, DietmarDietmarHarhoffHölzle, KatharinaKatharinaHölzleKaschub, VerenaVerenaKaschubMalagimani, SonalSonalMalagimaniMorgalla, UlrikeUlrikeMorgallaRose, RobertRobertRose2026-01-142026-01-142026-03https://publica.fraunhofer.de/handle/publica/503108https://doi.org/10.24406/publica-705110.1016/j.respol.2025.10538110.24406/publica-70512-s2.0-105022708208The integration of AI into scientific work holds significant potential to accelerate innovation. We surveyed researchers in two leading German research organizations to examine AI adoption, barriers, and perceived impact on research. Researchers are widely using AI tools - often for primary and creative tasks - and many expect the technology to be transformative for research. Effective use appears linked to both hands-on experience and engagement with learning resources. A persistent gender gap in AI use is closely associated with differences in familiarity, suggesting a potential focus for organizational efforts. Legal uncertainty and privacy concerns also emerge as major barriers, with researchers calling for clear, high-level regulatory guidance. Overall, our findings suggest directions where institutional actions might be explored to promote more equitable and effective AI adoption.entrueTechnology adoptionAIResearchInnovationSurvey evidencejournal article