# Public Perception of AI in Scientific Research New study reveals attitudes toward AI adoption in academic studies - **Source:** [Research Policy](https://www.sciencedirect.com/journal/research-policy) - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/450d3106-988e-41b6-a606-92caddb567fa ## Growing AI adoption in science The study examines increasing integration of AI tools across scientific disciplines, from literature reviews to data analysis. Researchers note AI's potential to accelerate discovery while raising questions about methodology transparency. Adoption rates vary significantly by field, with computer science leading and social sciences more cautious. This mirrors broader digital transformation trends in academia over the past decade. ## Public trust divides by application Survey results show 58% approval for AI assisting data processing but only 32% support for AI-generated hypotheses. Participants expressed strongest reservations about AI's role in peer review and research evaluation. The trust gap correlates with respondents' familiarity with scientific methods - laypeople show higher skepticism than STEM professionals. These patterns hold across North American and European samples. ## Transparency as key concern 73% of respondents demand clear disclosure when AI tools contribute to research outputs. Many participants worry about undisclosed AI involvement potentially compromising research integrity. The study identifies particular concern about generative AI's ability to produce plausible but unverified content. Researchers suggest standardized reporting frameworks could address these transparency issues. ## Demographic variations Younger respondents (18-35) show 22% higher AI acceptance than those over 55. Gender differences emerge in concerns about job displacement, with women expressing stronger worries. The study found no significant variation by education level once field of expertise was controlled for. Regional differences appear minimal in adjusted models. ## Field-specific acceptance Medical researchers report highest comfort with AI-assisted diagnostics (67% approval). Social scientists show greatest resistance, citing concerns about algorithmic bias in qualitative analysis. Physical sciences demonstrate polarized views - strong support for computational tasks but skepticism about theoretical contributions. These patterns suggest domain-specific adoption challenges. ## Institutional trust factors Respondents show 40% higher confidence in AI tools vetted by academic institutions versus commercial providers. University-affiliated researchers enjoy more public trust than industry scientists when deploying AI methods. The study highlights the importance of peer review processes in legitimizing AI-assisted research across disciplines. ## Future adoption barriers Researchers identify three key obstacles: lack of standardized evaluation metrics (62%), insufficient training (55%), and publication policies (48%). Early career scientists report feeling particularly unprepared for AI integration. The study notes rapid evolution of tools outpaces institutional capacity to establish best practices. ## Policy implications Findings suggest need for discipline-specific AI adoption guidelines and mandatory disclosure standards. Public acceptance appears contingent on demonstrated oversight mechanisms and error-correction protocols. The research team recommends phased integration approaches that maintain human oversight in critical research functions while leveraging AI's analytical strengths. ## Sources covering this story - https://www.sciencedirect.com/science/article/pii/S0048733326002234?dgcid=rss_sd_all — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._