Recent advances in artificial intelligence (AI) technology create a strong impetus for policymakers to draft regulations aimed at reducing risks while maximizing benefits. As jurisdictions around the world seek the right policy answers to the ethical, socioeconomic and democratic implications of AI, a wide range of stakeholders is engaging in the global policy debate that will inform policymaking and shape future policies on AI. Publishing AI ethics guidelines and policy documents is a prominent way for governmental, corporate, and civil society actors to participate in deliberative processes and agenda-setting, allowing them to shape perceptions, understandings, and assessments of the technology. To shed light on what is being said (and how) in these AI policy documents, this paper employs natural language processing (NLP) techniques and neural networks. It builds on an original large-scale data set of over 400 AI policy documents and a rich hand-coded sample to train deep-learning classifiers (via TensorFLow) that predict topics, sentiments, and types of statements at the sentence level. The resulting data allows us to paint a fine-grained picture of the global discourse on AI policy, unearthing interesting patterns in attention, tonality, and purpose of the documents.
Elia Soler Pastor (UPF)
Emotions shape important cognitive and social functions, such as processes of identity and relationship formation. Awe is a self-transcendent emotion that shifts our focus to an outward sense of universality and interconnectedness. Past research has experimentally studied the effects of awe on general connectedness to others, mostly through prototypical nature stimuli. Research on interpersonal awe and on its effects on connectedness to specific groups is scarce. Plus, we do not know whether awe has the power to bind people together even in challenging contexts that include instances of social rejection or discrimination, and more research is needed to understand this emotion in diverse, non-W.E.I.R.D. samples. We investigate the role of in a community sample of immigrants in Spain (N = 100). Participants were exposed (in counterbalanced order) to: (a) a nature awe video and (b) an interpersonal awe video (i.e., depicting a collective tradition of the host culture). In between the awe-inducing videos, they all visualized an immigrant rejection video that elicited a negative emotional response. Repeated measures’ analyses show that both awe-inducing videos made people feel strongly connected to humanity, and that cultural awe made them as well feel connected to the host culture. Moderation results show that awe felt from the cultural (vs. nature) experience triggered connectedness to both the host culture and humanity, even after the rejection experience. Finally, explorative regressions show that there may be important contextual predictors of awe that are related to connectedness and have not been explored in the past (i.e., identity and social network variables). Our findings suggest that self-transcendent emotions such as awe may play an important role in identity processes and in intergroup or intercultural relations.
Pau Vall-Prat (UB)
How does intra-elite competition influence electoral fraud? Most intra-elite competition literature focuses on electoral autocratic contexts in which different types of electoral fraud or electoral manipulation are prevalent. However, we lack explanations regarding to what extent intra-elite competition can influence informal practices of electoral fraud. In this paper I explore whether the presence of a regional elite party in electoral contests in the early 20th century changed the extent of electoral fraud. I analyze the political consequences of a new elite party in electoral competition through a combination of qualitative and anecdotal evidence, a wide array of electoral forensic techniques, and some causal inference methods. The results point out that the levels of electoral fraud decreased under the presence of new elites and that this was related to some shifts on the types and intensity of electoral manipulation. New regional elites, without direct access to the administrative apparatus, had to resort to electoral mobilization resources to prevent electoral fraud and to be electorally competitive. All in all, this paper shows that intra-elite competition can also foster informal democratization practices through the cleansing of certain electoral procedures.
Laura Gutiérrez Zárate (UAB)
In the past few years, non-hierarchical structures have emerged in women’s movements where leadership figures are not formally established. Under these structures, leadership is usually composed of a group of individuals that share responsibilities and resources in a non-authoritarian way. The scholarly literature has recognized movement leaders as crucial decision-makers, yet scholars have dedicated little attention to analyze how their individual characteristics influence the movement’s strategies, and by extension, their capacity to achieve policy outcomes. Leadership figures possess individual skills, expertise, and knowledge that influences their decisions and how they interact with other actors, including the internal group of leaders and external political actors. These individual characteristics have been theorized as leadership capital and categorized in three dimensions: cultural, social, and symbolic capital. To gain access to the policy-making process, women’s movements require building political alliances that will act as mediators between the movement and the institutions. Therefore, an important task for movement leaders is influencing these mediators or political actors so that may shape policies in their favor. Building on leadership capital, this study proposes to include a fourth category of analysis to understand how leadership capital characteristics interact under shared leadership structures in women’s movements, where internal coalitions need to be developed among the leadership to succeed. Through a systematic comparative case study set in the period 2016-2021, this study’s qualitative approach utilizes in-depth interviews to compare two women’s movements from Mexico: the Olimpia and the Anti-femicide movements. In the case studies, I will demonstrate that the leaders of the Olimpia movement were successful because they built an internal coalition that strengthened their leadership capital which helped them achieve the passage of a law through their political allies. On the contrary, the leaders of the Anti-femicide movement were not able to succeed due to missing characteristics of leadership capital and a failure to form a strong internal coalition. This paper thus advances leadership and social movement research by proposing a framework to study the components of leadership capital under a shared leadership structure which serves to analyze how they influence the movement’s capacity to achieve political outcomes.
Joan-Josep Vallbé (UB)
Eva Anduiza (UAB)
Abel Escribà-Folch (UPF)
Toni Rodon (UPF)
Ruben García del Horno (UAB)
Carmen Ramírez Folch (UB)
Jonathan Kent (UPF)
Marina Muñoz Puig (UPF)