Search for: All records

The COVID-19 pandemic demonstrated the importance of social distancing practices to stem the spread of the virus. However, compliance with public health guidelines was mixed. Understanding what factors are associated with differences in compliance can improve public health messaging since messages could be targeted and tailored to different population segments. We utilize Twitter data on social mobility during COVID-19 to reveal which populations practiced social distancing and what factors correlated with this practice. We analyze correlations between demographic and political affiliation with reductions in physical mobility measured by public geolocation tweets. We find significant differences in mobility reduction between these groups in the United States. We observe that males, Asian and Latinx individuals, older individuals, Democrats, and people from higher population density states exhibited larger reductions in movement. Furthermore, our study also unveils meaningful insights into the interactions between different groups. We hope these findings will provide evidence to support public health policy-making. 

In 2021 GW Engineering was awarded funding to launch an interdisciplinary program on trustworthy AI. Designing Trustworthy AI in Systems (or DTAIS) brings together PhD students from systems engineering and computer science to co-design research and tackle the conceptual and methodological bridge building that cross disciplinary work demands. This paper focuses on how this work has been accomplished thus far, in the context of the cornerstone summer incubator, and shares some of the lessons learned. The 10-week summer incubator course, which was designed specifically for this program, brings systems engineers and computer science PhD students to make sense of “AI in the wild” (real world settings) and build short-run research prototypes together. Leveraging the interdisciplinarity of the core program faculty, the group established a fertile middle ground where a mixed method ethos, design sprint rhythm and intentional sense of community enlivens the normative student-advisor modality most PhD students experience. Along the way, the definitional challenge of what is meant exactly by trust and trustworthiness within a particular problem domain and literature is given plenty of room to form, fall apart and form again through discussion, practice, and reflection. With two iterations of the summer incubator course to glean from, we report on the difficulties of rewiring student-advisor dynamics and the positive effects of growing a diverse community. This represents a potential roadmap for how to scaffold interdisciplinarity in engineering doctoral education. 

This study examines the relationship between online communication by the Proud Boys and their offline activities. We use a supervised machine learning model to analyze a novel dataset of Proud Boys Telegram messages, merged with US Crisis Monitor data of violent and nonviolent events in which group members participated over a 31-month period. Our analysis finds that intensifying expressions of grievances online predict participation in offline violence, whereas motivational appeals to group pride, morale, or solidarity share a reciprocal relationship with participation in offline events. This suggests a potential online messaging–offline action cycle, in which (a) nonviolent offline protests predict an increasing proportion of motivational messaging and (b) increases in the frequency and proportion of motivational appeals online, in turn, predict subsequent violent offline activities. Our findings offer useful theoretical insights for understanding the relationship between online speech and offline behavior. 

Background. Vaccine misinformation has been widely spread on social media, but attempts to combat it have not taken advantage of the attributes of social media platforms for health education. Methods. The objective was to test the efficacy of moderated social media discussions about COVID-19 vaccines in private Facebook groups. Unvaccinated U.S. adults were recruited using Amazon’s Mechanical Turk and randomized. In the intervention group, moderators posted two informational posts per day for 4 weeks and engaged in relationship-building interactions with group members. In the control group, participants received a referral to Facebook’s COVID-19 Information Center. Follow-up surveys with participants (N = 478) were conducted 6 weeks post-enrollment. Results. At 6 weeks follow-up, no differences were found in vaccination rates. Intervention participants were more likely to show improvements in their COVID-19 vaccination intentions (vs. stay same or decline) compared with control (p = .03). They also improved more in their intentions to encourage others to vaccinate for COVID-19. There were no differences in COVID-19 vaccine confidence or intentions between groups. General vaccine and responsibility to vaccinate were higher in the intervention compared with control. Most participants in the intervention group reported high levels of satisfaction. Participants engaged with content (e.g., commented, reacted) 11.8 times on average over the course of 4 weeks. Conclusions. Engaging with vaccine-hesitant individuals in private Facebook groups improved some COVID-19 vaccine-related beliefs and represents a promising strategy. 

Utilizing an original data set of public Telegram channels affiliated with a right-wing extremist group, the Proud Boys, we conduct an exploratory analysis of the structure and nature of the group’s presence on the platform. Our study considers the group’s growth, organizational structure, connectedness with other far-right and/or fringe factions, and the range of topics discussed on this alternative social media platform. The findings show that the Proud Boys have a notable presence on Telegram, with a discernable spike in activity coinciding with Facebook’s and Instagram’s 2018 deplatforming of associated pages and profiles with this and other extremist groups. Another sharp increase in activity is then precipitated by the attack on the U.S. Capitol Building on January 6, 2021. By February 2022, we identified 92 public Telegram channels explicitly affiliated with the Proud Boys, which constitute the core of a well-connected network with 131,953 subscribers. These channels, primarily from the United States, also include international presences in Australia, New Zealand, Canada, the UK, and Germany. Our data reveals substantialinteraction between the Proud Boys and other fringe and/or far-right communities on Telegram, including MAGA Trumpists, QAnon, COVID-19-related misinformation, and white-supremacist communities. Content analyses of this network highlights several prominent and recurring themes, including opposition to feminism and liberals, skepticism toward official information sources, and propagation of various conspiracy beliefs. This study offers the first systematic examination of the Proud Boys on Telegram, illuminating how a far-right extremist group leverages the latitude afforded by a relatively unregulated alternative social media platform. 

Influence operations are large-scale efforts to manipulate public opinion. The rapid detection and disruption of these operations is critical for healthy public discourse. Emergent AI technologies may enable novel operations that evade detection and influence public discourse on social media with greater scale, reach, and specificity. New methods of detection with inductive learning capacity will be needed to identify novel operations before they indelibly alter public opinion and events. To this end, we develop an inductive learning framework that: (1) determines content- and graph-based indicators that are not specific to any operation; (2) uses graph learning to encode abstract signatures of coordinated manipulation; and (3) evaluates generalization capacity by training and testing models across operations originating from Russia, China, and Iran. We find that this framework enables strong cross-operation generalization while also revealing salient indicators-illustrating a generic approach which directly complements transductive methodologies, thereby enhancing detection coverage. 

Background: Distrust and partisan identity are theorized to undermine health communications. We examined the role of these factors on the efficacy of discussion groups intended to promote vaccine uptake. Method: W e analyzed survey data from unvaccinated Facebook users (N = 371) living in the US between January and April 2022. Participants were randomly assigned to Facebook discussion groups (intervention) or referred to Facebook ’s COVID-19 Information Center (control). We used Analysis of Covariance to test if the intervention was more effective at changing vaccination intentions and beliefs compared to the control in subgroups based on participants ’ p artisan identity, political views, and information trust views. Results: W e found a significant interaction between the intervention and trust in public health institutions (PHIs) for improving intentions to vaccinate (P = .04), intentions to encourage others to vaccinate ( P = .03), and vaccine confidence beliefs ( P = .01). Among participants who trusted PHIs, those in the intervention had higher posttest intentions to vaccinate ( P = .008) and intentions to encourage others to vaccinate ( P = .002) compared to the control. Among non-conservatives, participants in the intervention had higher posttest intentions to vaccinate ( P = .048). The intervention was more effective at improving intentions to encourage others to vaccinate within the subgroups of Republicans ( P = .03), conservatives (P = .02), and participants who distrusted government ( P = .02). Conclusions: Facebook discussion groups were more effective for people who trusted PHIs and non-conservatives. Health communicators may need to segment health messaging and develop strategies around trust views. 

Abstract—We consider the ability of CLIP features to support text-driven image retrieval. Traditional image-based queries sometimes misalign with user intentions due to their focus on irrelevant image components. To overcome this, we explore the potential of text-based image retrieval, specifically using Contrastive Language-Image Pretraining (CLIP) models. CLIP models, trained on large datasets of image-caption pairs, offer a promising approach by allowing natural language descriptions for more targeted queries. We explore the effectiveness of textdriven image retrieval based on CLIP features by evaluating the image similarity for progressively more detailed queries. We find that there is a sweet-spot of detail in the text that gives best results and find that words describing the “tone” of a scene (such as messy, dingy) are quite important in maximizing text-image similarity. 

Online misinformation promotes distrust in science, undermines public health, and may drive civil unrest. During the coronavirus disease 2019 pandemic, Facebook—the world’s largest social media company—began to remove vaccine misinformation as a matter of policy. We evaluated the efficacy of these policies using a comparative interrupted time-series design. We found that Facebook removed some antivaccine content, but we did not observe decreases in overall engagement with antivaccine content. Provaccine content was also removed, and antivaccine content became more misinformative, more politically polarized, and more likely to be seen in users’ newsfeeds. We explain these findings as a consequence of Facebook’s system architecture, which provides substantial flexibility to motivated users who wish to disseminate misinformation through multiple channels. Facebook’s architecture may therefore afford antivaccine content producers several means to circumvent the intent of misinformation removal policies. 

Systems engineering has often concerned itself with how operator and customer roles change when systems change. In the context of automated vehicles (AVs), it has been assumed that operators will be removed from the system architecture; however, new insights reveal that the role of operators, typically thought of as drivers, has been transformed, not eliminated. In this study, we identify how different types of door-to-door transportation services use varying organizational architectures to achieve required functions, and explore how these architectures might this change with emergence of automated door-to-door transportation services. We draw on prior research, archival documents, and semi-structured interviews with AV technical and operational experts to identify and detail required functions for these services. Preliminary results reveal that, counter to the commonly-held belief, the structures of commercial AV services more closely parallel traditional taxi organizations rather than current ride-hailing services based on their capital cost and human labor requirements. Future research will explore short and long-term development pathways for AV systems and their associated structural and functional requirements. While the structures of these AV companies will continue to develop alongside the automation technologies, early explorations of AV organizations can reveal multiple possible development pathways for AV services and highlight potentially desirable or undesirable intermediary stages.