O presente artigo busca elucidar como os dados pessoais de usuários das mídias sociais são re- levantes para ferramentas de personalização da informação e como elas influenciam na frag- mentação social, podendo criar um ambiente virtual problemático para a finalidade demo- crática. O objeto desta pesquisa é a preservação do debate construído sobre informações e fatos para objetivos democráticos, uma vez que a cir- culação de informação nas redes sociais atinge seus usuários através de bolhas informacio- nais – ocasionando o menor acesso à informa- ção plural e contraditória. Serão abordados dois instrumentos para a ocorrência de tal fato: a ar- quitetura de controle, efetuada pelo trabalho algorítmico, e o micro direcionamento de pro- pagandas ou anúncios patrocinados. Ambas téc- nicas são alimentadas pela matéria-prima com- posta pelos dados pessoais dos usuários, tendo como objetivo a aproximação entre anunciante/ vendedor e potenciais consumidores – o que se torna um problema quando esse ambiente de consumo é utilizado para o contexto político-
-eleitoral. Com base em alguns exemplos prá- ticos, campanhas eleitorais dos Estados Unidos de 2016 e Brasil 2018 e 2022, busca-se proble- matizar as plataformas para compreender a so- berania política e do consumidor, com base na teoria de Cass Sunstein.

Since the coming into effect of the current national legislation on the subject, in 2005, judicial reorganizations have assumed great importance in Brazil. The considerations brought forward in this study will specifically address the issues confronted by firms when considering pleading for judicial reorganizations, reviewing the requirements, conditions and information entailed to initiate proceedings. Subsequently, the criteria for granting the processing of reorganizations by the judiciary will be reviewed, as well as the legal implications of such decisions. Once the supporting technical background is provided, critical remarks will be made on the Brazilian legal framework and market behavior towards undertaking reorganization procedures.

This paper presents additional theoretical, qualitative, and empirical evidence to understand the profiles of Brazilian citizens that share political fake news online and their potential motivations. The study introduces exclusive data collection through a national telephone survey, a taylormade focus group, and quantitative multivariate modeling. The qualitative exploration exposed fake news sharing motivations such as social approval, attention attraction, or strong feelings. The empirical results show that income level (especially Brazilian middle class), religious preferences (mostly evangelicals), and online frequency of exposure to fake news are key profile drivers for sharing fake news.

This book offers a synthesis of investigations on the ethics, governance and policies affecting the design, development and deployment of artificial intelligence (AI). Each chapter can be read independently, but the overall structure of the book provides a complementary and detailed understanding of some of the most pressing issues brought about by AI and digital innovation. Given its modular nature, it is a text suitable for readers who wish to gain a reliable orientation about the ethics of AI and for experts who wish to know more about specific areas of the current debate.

Disinformation, hate speech and political polarization are evident problems of the growing relevance of information and communication technologies (ICTs) in current societies. To address these issues, decision-makers and regulators worldwide discuss the role of digital platforms in content moderation and in curtailing harmful content produced by third parties. However, intermediary liability rules require a balance that avoids the risks arising from the circulation at scale of harmful content and the risks of censorship if excessive burdens force content providers to adopt a risk-averse posture in content moderation. This piece examines the trend of altering intermediary liability models to include ‘duty of care’ provisions, describing three models in Europe, North America and South America. We discuss how these models are being modified to include greater monitoring and takedown burdens on internet content providers. We conclude with a word of caution regarding this balance between censorship and freedom of expression.

Machine learning (ML) is widely used to moderate online content. Despite its scalability relative to human moderation, the use of ML introduces unique challenges to content moderation. One such challenge is predictive multiplicity: multiple competing models for content classification may perform equally well on average, yet assign conflicting predictions to the same content. This multiplicity can result from seemingly innocuous choices made during training, which do not meaningfully change the accuracy of the ML model, but can nevertheless change what the model gets wrong. We experimentally demonstrate how content moderation tools can arbitrarily classify samples as “toxic,” leading to arbitrary restrictions on speech. We use the principles set by the International Covenant on Civil and Political Rights (ICCPR), namely freedom of expression, non-discrimination, and procedural justice to interpret the effects of these findings in terms of Human Rights. We analyze (i) the extent of predictive multiplicity among popular state-of-the-art LLMs used for detecting “toxic” content; (ii) the disparate impact of this arbitrariness across social groups; and (iii) the magnitude of model multiplicity on content that is unanimously recognized as toxic by human annotators. Our findings indicate that the up-scaled algorithmic moderation risks legitimizing an “algorithmic leviathan”, where an algorithm disproportionately manages human rights. To mitigate such risks, our study underscores the need to identify and increase the transparency of arbitrariness in content moderation applications. Our findings have implications to content moderation and intermediary liability laws being discussed and passed in many countries, such as the Digital Services Act in the European Union, the Online Safety Act in the United Kingdom, and the recent TSE resolutions in Brazil.

No Brasil, existe uma preocupação crescente sobre o uso de propaganda computacional e a polarização política
que ela pode causar. Neste memorando, analisamos dados sobre notícias e informações políticas compartilhadas
no Twitter no período que antecedeu a eleição presidencial brasileira de 2018

Healthcare systems across the globe are struggling with increasing costs and worsening outcomes. This presents those responsible for overseeing healthcare with a challenge. Increasingly, policymakers, politicians, clinical entrepreneurs and computer and data scientists argue that a key part of the solution will be ‘Artificial Intelligence’ (AI) – particularly Machine Learning (ML). This argument stems not from the belief that all healthcare needs will soon be taken care of by “robot doctors.” Instead, it is an argument that rests on the classic counterfactual definition of AI as an umbrella term for a range of techniques that can be used to make machines complete tasks in a way that would be considered intelligent were they to be completed by a human. Automation of this nature could offer great opportunities for the improvement of healthcare services and ultimately patients’ health by significantly improving human clinical capabilities in diagnosis, drug discovery, epidemiology, personalised medicine, and operational efficiency. However, if these AI solutions are to be embedded in clinical practice, then at least three issues need to be considered: the technical possibilities and limitations; the ethical, regulatory and legal framework; and the governance framework. In this article, we report on the results of a systematic analysis designed to provide a clear overview of the second of these elements: the ethical, regulatory and legal framework. We find that ethical issues arise at six levels of abstraction (individual, interpersonal, group, institutional, sectoral, and societal) and can be categorised as epistemic, normative, or overarching. We conclude by stressing how important it is that the ethical challenges raised by implementing AI in healthcare settings are tackled proactively rather than reactively and map the key considerations for policymakers to each of the ethical concerns highlighted.

In Brazil, there are rising concerns over computational propaganda and the political polarization it may cause. In this data memo, we analyze data about political news and information shared over Twitter in the period leading up to the 2018 Brazilian presidential election. We find that: (1) Brazil’s political discourse on social media is highly partisan, with leading candidate Jair Bolsonaro dominating the conversation on Twitter; (2) accounts associated with Luiz Inácio Lula da Silva and Fernando Haddad hashtags show the highest level of high frequency tweeting; (3) Brazilian Twitter users are sharing more professional political content on Twitter than junk news — the highest proportion in all the elections we have studied; (4) while Bolsonaro supporters spread the widest range of known junk news sources, Lula da Silva and Haddad supporters accounted for the highest volume of shares.