Playing "in the Loop". New Human-Software Relations in Human Computation Systems and Their Impact on Spheres of Everyday Life

点击翻译按钮获取中文摘要

Today, algorithms are found to outperform humans in various fields, with their speed and accuracy far exceeding human abilities in incomparable manners. However, algorithms still fail on many comparatively basic problems that are easily solved by humans. Moreover, most of the algorithmic approaches to solving these problems are based on machine learning and require people “in the loop” to generate large amounts of training data for the algorithms. An approach in which humans and algorithms work together to tackle problems that neither of them can easily solve on their own is known as Human Computation. Within the realm of human computation, Human Computation Systems (HCS) delegate particular computational steps or tasks to humans. This emerging and increasingly important research area in computer science is especially successful in the area of citizen science (CS). Here, projects are usually conceived as computer games in which participants playfully generate surplus value both for a specific research project and for the algorithms. While HCS raise numerous social questions, as field of research they are almost unexplored in cultural anthropology. The proposed project is dedicated to this new research field and focuses on two aspects: On the one hand, it focuses on the investigation of HCS from a subject-related and moral perspective. This includes questions about how the systems are formed and what role human actors play within them. On the other hand, the project focuses on the impacts that HCS in CS-projects have on our understandings of the daily spheres of play, work and science, that function as important frames of reference for our self-understanding. The project draws on theoretical and conceptual perspectives from the field of digital anthropology and technology, the examination of (digital) ethics and the role of the subject, and the exploration of the meanings of the aforementioned spheres of the everyday life. Three case studies form the core of the project and will be investigated microanalytically. In order to investigate the interplay and mutual dependencies of the actors involved in HCS – developers, researchers, users and code – and thus the systems they have jointly formed, the approach is based on method triangulation and is multi-perspectival. By combining the perspectives of cultural anthropology and computer science this project aims at gaining an in-depth understanding of HCS and their implications on everyday life. The objective of the project is to contribute to the exploration of emerging human-machine interactions in the field of AI and also to further opening cultural anthropology for interdisciplinary research about these cross-thematic and interdisciplinary phenomena.