STELA: a community-centred approach to norm elicitation for AI alignment.

Value alignment, the process of ensuring that artificial intelligence (AI) systems are aligned with human values and goals, is a critical issue in AI research. Existing scholarship has mainly studied how to encode moral values into agents to guide their behaviour. Less attention has been given to the normative questions of whose values and norms AI systems should be aligned with, and how these choices should be made. To tackle these questions, this paper presents the STELA process (SocioTEchnical Language agent Alignment), a methodology resting on sociotechnical traditions of participatory, inclusive, and community-centred processes. For STELA, we conduct a series of deliberative discussions with four historically underrepresented groups in the United States in order to understand their diverse priorities and concerns when interacting with AI systems. The results of our research suggest that community-centred deliberation on the outputs of large language models is a valuable tool for eliciting latent normative perspectives directly from differently situated groups. In addition to having the potential to engender an inclusive process that is robust to the needs of communities, this methodology can provide rich contextual insights for AI alignment.

Mapping global variation in human mobility.

The geographic variation of human movement is largely unknown, mainly due to a lack of accurate and scalable data. Here we describe global human mobility patterns, aggregated from over 300 million smartphone users. The data cover nearly all countries and 65% of Earth's populated surface, including cross-border movements and international migration. This scale and coverage enable us to develop a globally comprehensive human movement typology. We quantify how human movement patterns vary across sociodemographic and environmental contexts and present international movement patterns across national borders. Fitting statistical models, we validate our data and find that human movement laws apply at 10 times shorter distances and movement declines 40% more rapidly in low-income settings. These results and data are made available to further understanding of the role of human movement in response to rapid demographic, economic and environmental changes.

"Coronavirus EXPLAINED": YouTube, COVID-19, and the Socio-Technical Mediation of Expertise.

Since the coronavirus outbreak, YouTube has altered its content moderation policies to surface more "authoritative information" while removing videos that contain "medically unsubstantiated claims." This was made urgent by incidents like a live-stream interview of renowned British conspiracy theorist David Icke-in which he falsely linked the spread of the coronavirus to 5G technology-that gained substantial traction online. Behind these events, however, lies a tension between the need for authoritative medical information and the socio-technical mediation that enables multiple, competing voices to lay claim to such authority on YouTube; a tension exacerbated by the current pandemic. Following an investigation into the sources and types of video content average users were likely to see when searching for information about the coronavirus on the site, we suggest that through its incentive structure and participatory affordances, YouTube may have subordinated expertise to a logic of likability-leaving institutional experts trailing behind.