This community service-based study responds to the urgency of strengthening digital literacy and awareness of hoaxes among elementary school students amid the increasing use of gadgets. The rapid penetration of digital technology has made children active internet users, but without adequate critical thinking skills, they remain vulnerable to false information and unethical digital behavior. This Program was held at SDN Pandean 1 and involved students in grades IV–VI through structured digital literacy socialization activities that focused on responsible use of gadgets, basic fact verification skills, and digital ethics. Using a descriptive qualitative approach, data was collected through observation, in-depth interviews with students and teachers, and documentation during program Implementation. The findings suggest that although students demonstrate high technical expertise in using digital devices, their ability to critically evaluate information, especially in identifying hoaxes, is still limited. However, such socialization activities contribute to increased awareness in verifying the source of information, recognizing misleading content, and adopting more ethical online behavior. This community service initiative highlights the importance of integrating digital literacy education into elementary learning environments. The results suggest that collaborative involvement between schools, teachers, and parents is essential to build students’ critical digital competencies from an early age.

The continued influence effect (CIE) refers to the persistent influence of False Information (FI), even after it has been corrected. This effect has been replicated in numerous experiments and carries significant societal implications. While the existing literature provides various explanations for the CIE, a comprehensive synthesis of these explanations remains lacking. This lack of synthesis hinders the ability to optimize recommendations for mitigating the effect and guiding future research. To address this, a conceptual reorganization of existing explanations is proposed within a heuristic framework encompassing three perspectives: cognitive, motivational, and rational. Drawing on this three-dimensional framework, recommendations are provided for mitigating this social issue and guiding future research.

The hallucination issues in generative artificial intelligence (GAI), stemming from training data bias and algorithmic misjudgment, constitute a significant source for information security problems. The resultant novel security threats, particularly the dissemination of false and biased information, pose a bottleneck constraining the development of GAI. Existing supervision mechanisms mostly focus on the binary game between the government and enterprises, neglecting the supervision value of users as direct risk-bearers. Based on the evolutionary game theory, this paper constructs a tripartite dynamic game model among enterprises, government and users, and analyzes the synergistic evolution mechanisms of responsible innovation in enterprises, user participation in supervision and the reward and punishment policy by government. Through the tripartite strategy matrix and replicator dynamics equation, we analyze the evolutionarily stable strategies and their influencing factors of each player, exploring the collaborative supervision mechanisms of GAI by numerical simulation. We find that: (1) There is an equilibrium point in the system, which is the evolutionarily stable strategies of responsible innovation in enterprises, user participation in supervision, and strict supervision by government. (2) The implementation of reward and punishment mechanisms by government for enterprises accelerates the realization of the evolutionarily stable strategies of the system. When the sum of government rewards obtained by enterprises and reputation premium brought by user feedback exceeds their additional costs, incentive measures such as subsidies or rewards adopted by the government are more effective in promoting enterprises to choose responsible innovation strategies, compared with simply intensifying punitive measures. (3) Designing an effective user evaluation feedback mechanism and increasing the rewards for user participation in supervision enhance the supervision willingness of users and accelerate the realization of evolutionarily stable states of collaborative supervision.

The European Union has established a rigorous system to control food imported from third countries and ensure consumer safety. However, the illegal trade in prohibited food items remains a significant issue. Non-compliant food, often counterfeited through false labeling, can bypass checks at border inspection points, choosing the least monitored entry points. These products pose a threat to public health, as they may contain products of animal origin, undeclared allergens, or come from unauthorized facilities. A recent example is the investigation initiated by the monitoring plan, which uncovered the illicit trade of food of animal origin in some ethnic shops. The Local Health Authority Naples 1 Center carried out extensive inspections of commercial establishments, finding that some foods labeled as wheat- or soy-based snacks were, in fact, prohibited food items from China. The Italian labels reported false information, hiding the real composition of the products. The operation led to the seizure of tons of food, including fish products and difficult-to-identify foods. The analysis detected undeclared allergens and, in some cases, traces of the African swine fever virus, although inactive. Following these discoveries, the Directorate General for Food Safety launched a program of coordinated controls at national level to combat illegal trade. This episode highlights the need to use advanced technologies and strengthen national and international collaboration to protect consumers.

ABSTRACT Social media influencers (SMIs) have become a prominent source of mental health – related information for young audiences, who often view them as role models and develop strong identification with them. However, SMIs sometimes share health content that is either misleading or overgeneralized, raising concerns about its potential effects. This study investigates the impact of such messaging and examines the role of identification with the message source. We conducted a pre-registered 3 (message type: misinformation vs. overgeneralized vs. control) × 2 (identification: higher vs. lower) between-subjects experiment with N = 549 Gen Z social media users in Germany recruited from an online panel, investigating three main outcomes: belief inaccuracy, behavioral intentions, and self-diagnostic tendencies. While neither misinformation nor overgeneralized messaging consistently affected behavioral intentions or self-diagnosis, findings showed that exposure to mental health-related misinformation by an SMI significantly increased belief inaccuracy, that is, a greater endorsement of inaccurate claims, regardless of whether participants strongly identified with the influencer.

ABSTRACT There is growing concern about the presence of false information on social media. Drawing on frameworks in the deception detection literature (e.g. the truth-default theory), the current research surveyed 536 individuals and asked them to recall a situation in which they detected information they believed to be false online. Results indicate that while participants often recalled politically relevant examples, gossip and product reviews were also common topics of false information discovery. A majority of individuals reported using after-the-fact evidence to determine the veracity of the content. However, this did not mean that most users explicitly conducted their own research. We discuss the role of false information in online spaces and connect it to the literature on social media and face-to-face deception.

The scientific literature is increasingly challenged by miscitation, publication bias, and the uncritical perpetuation of false information. One such case concerns the purported presence of caffeine in damiana, a botanical ingredient marketed for "sexual well-being". While consumer perception and regulatory implications make caffeine content highly relevant, a systematic review of the literature indicates that claims of caffeine in damiana (Turnera diffusa) largely stem from erroneous or circular referencing, including citations of studies where caffeine was never reported. To address this discrepancy experimentally, we conducted NMR analyses on authenticated aerial and seed material of T. diffusa and seeds of T. ulmifolia. No caffeine was detected, with a detection limit of < 4 nM, corresponding to < 4 µg caffeine per daily serving of damiana. These findings confirm that damiana is effectively caffeine-free. Beyond correcting a specific phytochemical misconception, this study highlights how flawed referencing practices can perpetuate modern "scientific myths", emphasizing the need for more rigorous citation ethics and the value of experimental data. Addressing such errors is critical not only for scientific accuracy but also for consumer trust and evidence-based regulatory decision-making.

The dissemination of false or inaccurate information, and its subsequent effect on behaviour, is not a new phenomenon. However, in recent years, alongside the emergence of multiple methods of information transmission, the speed, scale and volume of mis-, dis-, and malinformation (MDM) has reached a new threshold, contributing to societal polarisation and mistrust towards authorities. It has been postulated that the phenomenon may also play a role in the psychosocial process of radicalisation. By re-framing grievances or events and creating shared identities or networks amongst 'believers', exposure to MDM may interact with other audience vulnerabilities to facilitate radicalisation towards violent extremist narratives and networks. This systematic review will examine interventions designed to counter or reduce the effects of exposure to false information and their impact on violent extremism. The review will synthesise evidence across diverse methodologies and settings, identify gaps in the literature, and highlight best practices for reducing or countering the effects of false narratives and their subsequent influence on violent extremist behaviour and attitudes. The findings will inform policymakers, educators, and practitioners about actionable measures to combat the harmful consequences of false information in the context of violent extremism.

Artificial intelligence (AI) tools are increasingly promoted for clinical decision support in health care. While studies have assessed general-purpose AI chatbots on the accuracy and quality of clinical or drug-related questions, direct comparison among multiple clinically-focused AI chatbots is lacking. This study compared the quality of responses to drug information (DI) questions from multiple clinically-focused AI chatbots with responses from DI pharmacists. Thirty DI questions previously answered by DI pharmacy faculty were queried in four clinically-focused AI chatbots: OpenEvidence (OpenEvidence, Miami, FL), Clair (CaryHealth, Washington, DC), GlassHealth (GlassHealth, San Francisco, CA), and DougallGPT (Dougall Health, New York, NY). The general-purpose chatbot ChatGPT (Open AI, San Francisco, CA) was also queried. The quality of the responses was assessed independently by three pharmacy faculty members using the 25-point CLEAR scoring framework (completeness of content, lack of false information, evidence supporting the content, appropriateness, and relevance), a validated AI health information assessment tool. Mean total scores were categorized as either "poor" content (5-11 points on the CLEAR scale), "average" content (12-18 points), or "very good" content (19-25 points). Descriptive statistics of CLEAR scores were summarized, and Mann-Whitney U tests between faculty and AI-generated response scores were conducted. OpenEvidence had the highest mean total CLEAR score (17.82/25), followed by ChatGPT (15.72/25), both reflecting a content categorization of "average". Of the individual components of the CLEAR score, lack of false information consistently had the highest scores across AI chatbots, while evidence supporting the content had the lowest. When compared with faculty responses, all AI chatbots had statistically lower scores. While OpenEvidence had the highest overall CLEAR scores, few question responses from any chatbot were rated in the "very good" content category. Pharmacist expertise remains essential for ensuring answers to DI questions are of high quality.

Given that the spread of false information online often outpaces efforts to address it, it is important to understand i) what content is the most important to remove to prevent false beliefs and ii) whether lay intuitions about content removal are aligned with these findings. In four experiments, participants viewed social media posts containing false evidence for novel conspiracy theories. In Experiment 1 (N = 300), removing posts containing plausible evidence significantly decreased participants'conspiracy belief, while removing posts containing implausible evidence did not. In Experiments 2a-2c (Ns = 350, 352, 351, respectively), we investigated new participants' moderation preferences for the same posts. Worryingly, participants chose to remove posts containing implausible evidence more often than posts containing plausible evidence (Exp. 2a-2b). And, even when oriented to the goal of reducing belief in the broader conspiracy theory, participants still removed both types of posts at similar rates (Exp. 2c). Our results suggest that relying on lay intuitions to guide content removal decisions may result in prioritizing the removal of content that does not substantively contribute to the spread of broader false beliefs over content that does. These findings have important implications for moderation efforts and highlight the need to look beyond belief in isolated claims when developing effective misinformation interventions.

ABSTRACT The swift integration of generative Artificial Intelligence (AI) into automated, customer-facing, analytical, and strategic systems has sparked significant concern regarding the issue of “hallucinations” or false information generated by AI with high confidence. While previous work primarily identifies underlying causes of hallucinations related to model and training data quality, the work presented in this paper explores related implications for effective governance, auditing, and enterprise risk management. AI hallucinations affect operational, financial, legal, and strategic dimensions of an organization and impact accountability and responsibility. The paper proposes an integrated and interdisciplinary framework for effective governance and auditing of generative AI systems within an organization. Corresponding governance controls, AI assurance, auditing practices, cybersecurity, explainability, and compliance verification practices, as well as the necessary oversight by humans, are discussed to identify and mitigate the risks associated with hallucinations. By addressing issues of governance, audit, cybersecurity, and compliance, the paper contributes to responsible AI research and highlights the organizational and regulatory responsibilities associated with AI adoption at the enterprise level.

The use of large language models (LLMs) is increasingly common. However, LLMs may exhibit sycophancy, echoing users' beliefs while avoiding contradiction. In the present study, we describe sycophancy in general-purpose LLMs when applied to orthopaedic contexts. We investigated sycophancy in 2 general-purpose LLMs. We evaluated performance on 3 tasks: (1) accuracy on benchmark answering: LLMs were tested on validated benchmark orthopaedic questions, with correct and incorrect cues, and the change in accuracy and sycophancy error rate were determined; (2) user belief agreement: LLMs were provided with ambiguous statements and a user belief, and LLM agreement, contradiction, and uncertainty were described; and (3) false information detection: false information was placed within a task prompt to measure noncontradiction and propagation rates. Baseline factual accuracy on benchmark questioning was 78%, decreasing with correct hints (71%) (p = 0.49). With incorrect hints, LLM accuracy declined significantly (48%) (p < 0.001), with a sycophancy error rate of 52%. Presented with user beliefs about an indefinite, controversial statement, models echoed user beliefs in 56%, expressed uncertainty in 12%, and contradicted users in 32% of statements. In noncontradiction tasks, models perpetuated incorrect attributions 99% of the time yet reliably corrected statistical distortions 97% of the time. Although popular general-purpose LLMs have useful orthopaedic applications, they exhibit sycophancy, with a tendency toward agreement and without recognition of ambiguity. This is a key weakness to be addressed. Findings should be interpreted cautiously given the variability in model design, prompting, and models evaluated. The tendency of general-purpose LLMs to agree without recognizing clinical ambiguity may limit their reliability in orthopaedic applications.

Traditional survey methods make it difficult to collect sensitive data, as respondents provide false information or deliberately refuse to answer. The current investigation presents innovative methods utilizing multi-stage randomized response models (MRDRMs) to tackle challenges in precisely quantifying sensitive numerical variables while safeguarding respondent anonymity. The MRDRMs framework, comprising two as well as three-stage mathematical models. To employ two randomized response mechanisms designed to improve privacy protection and estimation efficiency to address perceptions like social desirability and over-estimation, which are prevalent in sensitive public health data collection. The main problem addressed in this study is the difficulty of obtaining precise estimates of sensitive quantitative variables because respondents often hesitate to provide truthful answers due to privacy concerns, fear, and social desirability bias. The incorporation of substantial randomization stages enables the models to provide unbiased and precise estimates of means and sensitivity levels, thereby ensuring the reliability of the data while reducing the psychological strain on participants. Theoretical analysis, particularly simulations conducted using Monte Carlo methods, suggests that MRDRMs, in both two-stage and three-stage formats, significantly improve the precision and relative efficiency of estimates compared to conventional randomized response strategies. The empirical validation carried out via a cross-sectional survey in the districts of Faisalabad, Lahore, Multan, and Rawalpindi in Punjab, Pakistan, examined not sufficiently reported COVID-19 cases along with vaccine hesitancy, thereby affirming the practical significance of the MRDRMs methods. The findings demonstrate a substantial improvement in estimation accuracy accompanied by reduced variance compared to conventional approaches. The results highlighting the effectiveness of MRDRM-I and MRDRM-II as robust methods for sensitive data collection. The findings from this research highlight the potential of MRDRMs for enhancing the acquisition and evaluation regarding sensitive data, strengthening consistent, ethical, and tangible results in research related to public health. These improvements strengthen privacy protection and enhance the validity of responses in sensitive surveys, offering meaningful contributions to public health monitoring and evidence-based policy development at the global level. The proposed models can be applied in public health surveillance, epidemiological studies, social science research, and policy-sensitive domains requiring accurate and privacy-preserving data collection.

ABSTRACT The United States is experiencing an influx of executive orders (EOs) as well as proposed and enacted laws restricting how institutions and educators attend to diversity, equity, and inclusion (DEI). We investigated the influence of the contentious sociopolitical climate created by anti-DEI laws in the context of higher education biology. Accurate and inclusive approaches of addressing the biological and social dimensions of diverse human experiences is becoming ever more imperative, particularly because of the way inaccurate claims about biology are being wielded to justify prejudicial EOs and legislation. Through the lens of faculty agency, we examined how biology instructors’ agentic actions (behavior aimed at achieving one’s goals) and agentic perspectives (viewing the self as capable of achieving those goals) are being influenced by anti-DEI laws. Participants expressed minimally changing their instructional content, although they became more aware of possible negative consequences of teaching potentially controversial topics. This led to participants making subtle changes to avoid scrutiny regarding their teaching. We call for more coalition building across administrators, faculty, and students for ongoing and proactive resistance against the anti-DEI movement.

Readers today face increasing challenges in distinguishing genuine news from misinformation and identifying biased content within algorithmically curated media environments. While prior research has focused on automated detection and adjustments to news recommendation systems, less attention has been given to user-facing interventions that shape how individuals engage with news. To address this gap, we introduce a browser-based personal informatics dashboard designed to support online news consumption by increasing awareness of bias, source quality, and credibility. In a between-subjects study (n = 100), our tool improved users’ ability to identify false information and encouraged more active engagement with news content (p < 0.05). We also observed increased lateral reading behaviors, such as cross-checking sources, suggesting deeper critical engagement. Findings are based on an integrated set of intervention features and combine statistical analyses with exploratory insights into user behavior. By adopting a personal informatics perspective, this work shifts focus from passive misinformation detection toward active, user-centered media literacy and more reflective interaction with news.

ABSTRACT Misinformation poses an important challenge for democracy, yet political debates appear to have polarized citizens’ perceptions and responses to misinformation. In a first experiment administered during the 2021 Canadian federal election, respondents were asked how acceptable it is for politicians to spread false information on climate change and COVID-19 for randomly assigned purposes. While most citizens considered misinformation unacceptable, opposing government action on these issues and having a right-wing ideology were associated with greater tolerance. I investigate these ideological differences further using observational and experimental data collected during the 2022 Quebec provincial election. I demonstrate that right-wing citizens are less likely to believe in the existence of an impartial truth, more indifferent to misinformation, and less supportive of measures to combat it than left-wing citizens. These patterns may partly reflect perceptions on the right that public discourses about misinformation are biased against them. The findings suggest that combatting misinformation has become a negatively valenced concept on the right and highlight how the politicization of misinformation debates can influence our ability to respond to it.

BackgroundSocial media has become a central channel for the dissemination of health information, enabling rapid sharing of evidence-based guidance while also facilitating the spread of inaccurate or misleading content. Exposure to such information has been associated with changes in vaccination-related attitudes and decision-making. These effects may not occur uniformly across populations, and there is limited understanding of how health information circulates within Spanish-language social media networks or which communication strategies are most effective in promoting engagement and informed decision-making. This study describes a protocol to evaluate a community-engaged intervention designed to address these gaps.MethodsThis study will use a non-randomized stepped-wedge design to evaluate the implementation and effectiveness of the Dime La VerDAD intervention across community-based cohorts of promotores de salud in Chicago. Promotores will be grouped into geographically defined clusters and will transition from control to intervention at six-month intervals, such that all clusters receive the intervention by study end. The intervention will consist of a structured, bilingual training program focused on identifying inaccurate health claims, evaluating source credibility, and developing accessible, evidence-based social media content, including narrative-based messaging. Outcomes will be assessed using surveys, focus groups, social media analytics, and publicly available epidemiologic data. Primary and secondary outcomes will include changes in knowledge, communication practices, engagement with social media content, and vaccination-related decision-making. Analyses will use mixed-effects models to evaluate changes over time while accounting for clustering and repeated measures.DiscussionThis study will generate evidence on how health information is shared and interpreted within Spanish-language social media networks and evaluate whether a structured, community-engaged communication intervention improves the quality and reach of health messaging. Findings will inform the development of scalable, community-based strategies to support dissemination of reliable health information and promote informed decision-making in diverse populations.Trial registration:ClinicalTrials.govNCT06417762. Recruitment has begun and is ongoing at the time of manuscript submission.

Risks and Harms of Conversational Artificial Intelligence (CAI) Chatbot Use Among US Youth.

Faster initial retrieval of misinformation corrections predicts better long-term memory: An ERP study.

A Delay Observer for False Information Detection in Line Current Differential Protection