With the widespread use of AI-generated content (AIGC) on social media, its potential to spread misinformation poses threats to the public. Although AI disclosure is widely promoted as a transparency measure to prompt critical evaluation, its effectiveness in science communication remains controversial. This study conducted a within-subjects experiment (N = 433) to examine how AI disclosure affects perceived credibility of science communication texts and the moderating roles of readers' negative attitudes towards AI and audience involvement. The experiment manipulated AI disclosure labels and information veracity. The results revealed a truth-falsity crossover effect: AI disclosure significantly reduced the perceived credibility of correct information while unexpectedly increasing the perceived credibility of misinformation. Negative attitudes towards AI significantly moderated these effects, whereas audience involvement exerted only limited influence. These findings highlight the complex and sometimes counterproductive consequences of AI disclosure in science communication and suggest implications for cue-based processing, algorithm aversion, and the design of disclosure mechanisms.

The digitalisation of science communication has been widely promoted within the Open Science movement in Europe to foster the social impact of research, as well as a more participatory culture of science. Using semi-structured interviews, we explore Spanish women scientists' values and perceptions regarding digital science communication. Results highlight the social value of science communication as well as intrinsic motivation as factors to actively engage in disseminating, educating and promoting science digitally. Adopting Open Science principles, participants craft open access multimodal materials (e.g., educational short videos, podcasts), use supporting multimodal resources and digital tools, and engage in social media to reach broad audiences. Finally, we propose some policy recommendations and pedagogical guidelines in terms of digital literacy, digital genres, and science accommodation strategies to promote digital science communication.

TikTok has become a popular platform for science communication, particularly among younger audiences, allowing creators to reach broader audiences. However, concerns about the accuracy of science content shared on the platform have emerged, prompting this study to investigate the reliability of informal science communication by popular creators. Informal science communication is the casual sharing of scientific information on platforms like TikTok. The main objective is to assess how well this content adheres to established scientific principles and avoids misinformation. By analysing videos from creators with significant followings, we will evaluate their adherence to scientific accuracy and identify factors that influence it, such as the creators' backgrounds and platform algorithms. The findings will highlight trends in the accuracy of content, with some creators producing reliable information while others risk spreading misinformation. Ultimately, the research will provide recommendations for enhancing the accuracy of science content on TikTok, promoting critical thinking among viewers, and advancing informed science communication on social media.

This study investigates how Germany’s anti-lockdown and anti-vaccine protest movement, led mainly by the Querdenken network, allied with conspiracist and far-right groups, utilized scientific authority while opposing COVID-19 policy. We analyse posts published in 161 public Telegram channels using a computational pipeline that combines named-entity recognition, structural topic modeling, a BERT sentiment classifier, and an open-source large language model, Mixtral. We report that mentions of scientific information surged during periods of heightened policy uncertainty (e.g., national lockdowns and the vaccine-mandate debate), indicating tactical appeals to epistemic authority. References to science were initially scarce rather than hostile, but evolved into a selective, strategic endorsement: protest communities increasingly cherry-picked scientific claims to delegitimize containment measures (foremost, vaccination) while sidelining evidence contradicting their narrative. The findings show that, even among actors who reject official institutions, appeals to scientific language are strategically deployed as a discursive resource.

Scientists and experts using social media platforms to engage with the public risk negative public feedback, potentially harming their efforts. This paper addresses how negative user comments affect experts’ trustworthiness and the messages’ credibility depending on whether they frame their message as scientific versus anecdotal using an online study with a 2 (evidence type: scientific vs. anecdotal) x 3 (comments: neutral, negative-factual, negative-emotional) between-subjects design. The results suggest that relying on scientific evidence when engaging in emotionally charged discourses is beneficial. Negative-emotional comments have a significant negative impact on trustworthiness, which is especially pronounced when using anecdotal evidence.

Two studies, using real-time response measurement and interviews, explore how German recipients assess the trustworthiness of a voice-based communicative AI conveying science-related information with empathic and humorous expressions. In both a laboratory and an online study, humor was associated with short-term declines in trustworthiness, reflecting cultural expectations and appreciation of objectivity and neutrality. In contrast, empathic expressions are rated more trustworthy, but evoke less conspicuous effects. Cluster analysis identified four distinct patterns of evaluation, with two groups largely unaffected by affective cues and two skeptical of humor, underscoring the importance of personalization and adaptation in designing trustworthy communicative AI for science communication.

This paper uses the Aristotelian concept of pathos to investigate how scientists can persuade their audiences through emotional appeals in science crowdfunding videos (SCVs). SCVs are short videos created to promote and fund a research project through online crowdfunding platforms, and represent an emerging genre of science communication that connects experts and audiences. By adopting a linguistic and semiotic approach, a sample of 50 SCVs was analysed with qualitative analysis software to identify linguistic and non-linguistic resources that could appeal to viewers' emotions. The findings show a strong emphasis on positive emotional appeals, particularly strategies that foster kindness and friendship between scientists and audiences. In contrast, appeals to fear and pity were minimal, suggesting that SCVs focus mostly on building trust and empowering potential donors rather than evoking urgency or guilt. These results shed light on the role of emotions in science communication, particularly in the context of research funding.

While social media has been praised for youth engagement with science, evidence of its impacts remains fragmented. This scoping review reports on the impacts of social-media-based science communication on young audiences. A PRISMA-guided database search yielded 2,257 articles, which were screened to include only empirical articles studying social media’s behavioral, attitudinal, and cognitive impacts on audiences, including youth, in science or health contexts. Using Directed Qualitative Content Analysis, the impacts desired, measured, and observed were categorized in the 35 remaining articles. The most desired and measured impact was knowledge gain, while the most observed outcomes were interest and trust in science. Many studies desired specific impacts but failed to measure them. Impactful content was relevant, visually appealing, and emotionally engaging. However, studies recognized that unreliable actors may also manipulate these characteristics to spread misinformation. While many science communicators assume the importance of social-media-based science communication for young audiences, evidence of observed outcomes is limited and specific to platforms and topics. 

This study explores how YouTube content creators integrate scientific evidence into their videos by analyzing citation patterns across disciplines. The role of other alternative metrics is also considered. We apply Principal Component Analysis (PCA) to compare the citation count of 12,005 research articles from Biotechnology, Psychology, Astrophysics, and Ecology published between 2014 and 2023, including citations sourced from YouTube videos. Our findings provide a characterization of two principal components in evidence citation employed by various science communication stakeholders. The first component enhances a paper's visibility by driving social attention, while the second focuses on its social influence and impact, determined by the paper's quality and scientific relevance.