Warning people about the risk of AI error mitigates human acquisition of AI bias.

Although interventions that warn about the harms of misinformation may be effective in lowering the credibility of false information, they may also cause suspicion related to factually accurate information. To explore these contradictory outcomes of exposure to media literacy messages, we used an online survey experiment with a diverse sample of 1105 participants in the U.S. In this experiment, we randomly exposed people to traditional warning messages about the threats of misinformation or relativising warning messages that placed misinformation’s threat in the context of the abundance of honest information. We additionally varied the specificity of the warning message (i.e. topic specific versus generic). We did not find direct overall effects on truth discernment but observed conditional effects on decreasing beliefs in misinformation and negative spillover effects on the truth rating of accurate information. We conclude that the effectiveness of media literacy interventions is far from straightforward, and document how pre-existing media trust plays a key role in the effects of such interventions. Based on our findings, we suggest that the effectiveness of specific and general media literacy messages may be contingent upon tailoring the message to levels of existing (dis)trust.

To ensure the security of mobile payments, anti-fraud warning messages serve as a critical defensive interface between users and potential risks. The effectiveness of their design directly influences users' risk perceptions and security-related behaviors. The present study employed eye-tracking technology to examine the effectiveness of warning messages in mobile payment transfer scenarios and the impact of specific warning design features on user decision-making. Experiment 1 utilized a 2 (warning message: present vs. absent) × 3 (potential risk level: high, medium, low) within-subject design to test the fundamental role of warning message presence. Results indicated that the presence of warning messages significantly prolonged participants' reaction times when selecting the transfer option, suggesting a more cautious decision-making process. Building on Experiment 1, Experiment 2 employed a 2 (warning color: red vs. blue) × 2 (warning semantic type: imperative vs. reminder) × 3 (potential risk level: high, medium, low) within-subject design and incorporated eye-tracking technology to investigate the effects of these design variables and underlying attentional mechanisms. Red warnings and imperative semantics were both found to significantly increase the likelihood of transfer rejection, with these design advantages being particularly pronounced in high-risk contexts. These findings provide empirical evidence to guide mobile payment platforms in optimizing dynamically adaptive, context-sensitive anti-fraud warning designs.

Short text emergency warnings in high-risk settings: Behavioural and representational insights from the British high commission (BHC) in Nairobi, Kenya

Statement of PurposeFramed safety messages (gain- or loss- framed) can counteract the increase in risk taking that occurs when children are in a heightened positive mood. In previous research, framed...

INTRODUCTIONAlthough a substantial body of research has analyzed the effectiveness of cigarette package warning labels in tobacco control, the very general health warnings messages (HWMs) on cigarette packaging in China have shown limited effectiveness in deterring youth from smoking. Therefore, this study investigates the impact of specific and more detailed warning text messages on Chinese young people’s risk perception of smoking and their intention to quit.METHODSWe employed a randomized survey experiment to examine the impact of specific text-based warning labels on Chinese young people’s risk perception of smoking and intention to quit. The total effective sample size was 1064 participants. The subjects were divided into three groups: the first group served as the control group, which was shown the existing cigarette package warning labels; the second group was shown cigarette package warning labels related to cardiovascular, digestive, and respiratory diseases; and the third group was shown cigarette package warning labels related to sexual dysfunction.RESULTSThe respiratory disease-related warnings significantly increased young people’s awareness of smoking-related respiratory risks (p<0.01). The impact of warning labels for the three common diseases on enhancing young people’s overall risk perception of smoking (p<0.05) and their intention to quit exhibited only weak statistical significance (p<0.05). In contrast, warning labels related to sexual dysfunction significantly increased young people’s risk perception of smoking (p<0.001) and their intention to quit (p<0.001), with a much higher level of statistical significance compared to those related to the other three common diseases.CONCLUSIONSDetailed descriptions of the risks associated with all four diseases were positively correlated with awareness of smoking-related harm and the intention to quit. However, warnings related to sexual dysfunction had a greater level of statistical significance compared to those related to the other three common diseases. This stronger significance may be attributed to young people’s heightened concern about sexual dysfunction.

Background: There is a lack of empirical data on lay attitudes toward different sorts of deception in medicine. However, lay attitudes toward deception should be taken into account when we consider whether deception is ever permissible in a medical context. The objective of this study was to examine lay attitudes of U.S. citizens toward different sorts of deception across different medical contexts. Methods: A one-time online survey was administered to U.S. users of the Amazon “Mechanical Turk” website. Participants were asked to answer questions regarding a series of vignettes depicting different sorts of deception in medical care, as well as a question regarding their general attitudes toward truth-telling. Results: Of the 200 respondents, the majority found the use of placebos in different contexts to be acceptable following partial disclosure but found it to be unacceptable if it involved outright lying. Also, 55.5% of respondents supported the use of sham surgery in clinical research, although 55% claimed that it would be unacceptable to deceive patients in this research, even if this would improve the quality of the data from the study. Respondents supported fully informing patients about distressing medical information in different contexts, especially when the patient is suffering from a chronic condition. In addition, 42.5% of respondents believed that it is worse to deceive someone by providing the person with false information than it is to do so by giving the person true information that is likely to lead them to form a false belief, without telling them other important information that shows it to be false. However, 41.5% believed that the two methods of deception were morally equivalent. Conclusions: Respondents believed that some forms of deception were acceptable in some circumstances. While the majority of our respondents opposed outright lying in medical contexts, they were prepared to support partial disclosure and the use of placebos when it is in the patient's interests or when it is what the person would want. These results support the position that physicians should be allowed a greater degree of authority to make a professional judgment about whether deception might be morally warranted by the circumstances, provided that it doesn't involve outright lying.

Telemedicine use has risen significantly since the COVID-19 pandemic. Evidence suggests that the quality of care in telemedicine could be as good as in-person care, but this is likely context-dependent. Expert guidelines have declared the appropriate medical conditions, but often without empirical evidence that grapples with the fundamental information limitations facing telemedicine. We draw on the task-technology fit theory and empirical evidence around human communication to examine how the medical and social contexts affect the efficiency and clinical quality of primary care. We will use a population-based dataset from the Canadian province of British Columbia (BC) to inform a quasi-experimental study using propensity score matching (PSM). The treatment group will consist of telemedicine visits from April to December 2022. We will use PSM to create a control group of matched, in-person visits in the same period. We will then use cluster-robust linear regression to identify how specific medical conditions and social contexts are associated with higher rates of prescription, follow-up with primary care providers, emergency department visits and acute care admissions. We plan for the study to take place from 1 August 2025 to 1 August 2026. The Research Ethics BC has granted approval for this study (H21-02244-A006). Our findings will be shared with patients, healthcare providers and policymakers and disseminated through conference presentations and peer-reviewed publications.

The federated learning (FL) paradigm is well-suited for the field of medical image analysis, as it can effectively cope with machine learning on isolated multi-center data while protecting the privacy of participating parties. However, current research on optimization algorithms in FL often focuses on limited datasets and scenarios, primarily centered around natural images, with insufficient comparative experiments in medical contexts. In this work, we conduct a comprehensive evaluation of several state-of-the-art FL algorithms in the context of medical imaging. We conduct a fair comparison of classification models trained using various FL algorithms across multiple medical imaging datasets. Additionally, we evaluate system performance metrics, such as communication cost and computational efficiency, while considering different FL architectures. Our findings show that medical imaging datasets pose substantial challenges for current FL optimization algorithms. No single algorithm consistently delivers optimal performance across all medical FL scenarios, and many optimization algorithms may under-perform when applied to these datasets. Our experiments provide a benchmark and guidance for future research and application of FL in medical imaging contexts. Furthermore, we propose an efficient and robust method that combines generative techniques using denoising diffusion probabilistic models with label smoothing to augment datasets, widely enhancing the performance of FL on classification tasks across various medical imaging datasets.

Chapter 9 - Defining problematic use of medicinal cannabis: Theoretical and empirical considerations

Researchers are investigating a broad spectrum of factors affecting positively and/or negatively the evacuation decision-making process occurring after people at risk receive cyclone warnings and advisories. Previous studies suggest that early warnings themselves do not propagate evacuation processes to be investigated but, rather, that human risk perceptions do so. This in turn encourages the sociopsychological dimensions of risk perception to be evaluated, which must be done within a country’s own cultural context. In applying content analysis here, we review the literature on evacuation decision-making processes during rapidonset hazards, i.e., tropical cyclones, in coastal Bangladesh. We focus on three broad overlapping themes – early warning, risk perception, and evacuation decision-making. Major content-analysis findings suggest that two things – a lack of credibility in early warning messages and an inefficient dissemination process – tend to affect the risk perception of people at risk and are likely to eventually determine the success of evacuation decision-making. Findings also show that different socioeconomic and socio-cultural issues related to risk perception appear to be more influential than formal warning messages in propagating decisions to evacuate during a cyclone. Based on these results, we suggest specific policy recommendations for improving local evacuation efficiency.

Research has begun to explore the use of virtual humans (VHs) in medical interviews. When designed as supportive and "safe" interaction partners, VHs may improve such screenings by encouraging patients to disclose more personal information [2-3]. In medical contexts, patients often feel resistance to self-disclosure and engage in impression management to be viewed more positively by healthcare providers. This paper provides the first empirical evidence that VHs can reduce such resistance and impression management. In the context of health-screening interviews, we report a study in which participants interacted with a VH that was either teleo-operated by humans (Wizard-of-Oz) or fully-automated (AI). Independently, we manipulated whether participants believed the VH was controlled by humans or automation. As predicted, participants who believed they were interacting with a computer reported lower resistance to self-disclosure, lower impression management and higher system usability than those who believed they were interacting with a human operator. Whether the virtual human was actually operated by a human or AI only affected ratings of the system's usability. These results suggest that automated VHs can help overcome a significant barrier to obtaining truthful patient information in medical domains.

Research has begun to explore the use of virtual humans (VHs) in medical interviews. When designed as supportive and safe interaction partners, VHs may improve such screenings by encouraging patients to disclose more personal information [2-3]. In medical contexts, patients often feel resistance to self-disclosure and engage in impression management to be viewed more positively by healthcare providers. This paper provides the first empirical evidence that VHs can reduce such resistance and impression management. In the context of health-screening interviews, we report a study in which participants interacted with a VH that was either teleo-operated by humans (Wizard-of-Oz) or fully-automated (AI). Independently, we manipulated whether participants believed the VH was controlled by humans or automation. As predicted, participants who believed they were interacting with a computer reported lower resistance to self-disclosure, lower impression management and higher system usability than those who believed they were interacting with a human operator. Whether the virtual human was actually operated by a human or AI only affected ratings of the system's usability. These results suggest that automated VHs can help overcome a significant barrier to obtaining truthful patient information in medical domains.

This column presents preliminary findings of an intervention to support shared decision making in psychopharmacology consultation. The waiting area in an urban psychiatric medication clinic was transformed into a peer-run Decision Support Center featuring a user-friendly, Internet-based software program with which clients could create a one-page computer-generated report for use in the medication consultation. The Decision Support Center was used 662 times by 189 unique users from a young-adult and general adult case management team from October 2006 to September 2007. All clients had severe mental disorders. Only ten clients refused to use the intervention at some point during the pilot study. Focus groups with medical staff (N=4), clients (N=16), case managers (N=14), and peer-specialist staff (N=3) reported that the intervention helped to create efficiencies in the consultation and empower clients to become more involved in treatment-related decision making. A randomized controlled trial is currently in process.

Background:The COVID-19 pandemic had a major impact on public health, especially for smokers, who faced higher infection risks and more severe disease progression.This study aimed to investigate how public health messages influence smokers' perceptions of tobacco use.Methods: This randomized controlled trial involved 1,303 adult smokers allocated to one of the following three groups: (1) a control group receiving a general smoking risk message, (2) a treatment group receiving a loss-framed message emphasizing smoking risks during COVID-19, or (3) a treatment group receiving a gain-framed message highlighting the health benefits of quitting during COVID-19.Immediately after reading the message, participants completed an Implicit Association Test to measure their unconscious attitudes toward tobacco by focusing on positive and negative perceptions.Results: COVID-19-related messages had a greater impact on smokers' perceptions than general messages about smoking risks, with differences depending on age and level of nicotine dependence.The gain-framed message was particularly successful in reinforcing negative views of tobacco among older smokers.Both gain-and loss-framed messages proved especially effective for individuals with high nicotine dependence.Conclusion: Message framing plays a crucial role in shaping smokers' perceptions during respiratory disease outbreaks, such as COVID-19.For effective smoking cessation policies, messaging strategies should be tailored to individual characteristics such as age and nicotine dependence.This study provides empirical evidence supporting effective public health communication strategies, particularly in the context of a pandemic.

In the classical "jacket-calculator" dilemma it is postulated decision is regulated by a mental topical accounting process which orients people to consider a discount price when purchasing items. We proposed an adapted version of the classical "jacket-calculator" task re-framing the choice in a medical context. Our results supported the view that simple minimal mental accounts influence evaluation and choice in the medical context where time, instead of price, represents a fundamental cue of the decision analysis. The decision process adopted by participants tended to be associated with a lexicographic decision mechanism where "time" appears the most effective cue of a "take-the-best" heuristic to predict people's behaviour accurately. These findings broaden the body of evidence indicating that bounded rationality in human decisions is intrinsically connected with the decisional context and different contexts may elicit different mental accounting strategies. In addition, the study stressed the need to enhance the dialogue between the more recent paradigm of the ecological rationality with the classical interpretations of bounded rationality because the two paradigms, not rarely opposed to each other, can provide hints to the interpretation of the decision process, with practical considerations for future interventions in health education and public health.