9607 research outputs found
Sort by
Hannah Devinney. Gender and Representation: Investigations of Bias in Natural Language Processing
No full textIllegal loot box advertising on social media? An empirical study using the Meta and TikTok ad transparency repositories
Get PDFLoot boxes are gambling-like products inside video games that can be bought with real-world money to obtain random rewards. They are widely available to children, and stakeholders are concerned about potential harms, e.g., overspending. UK advertising must disclose, if relevant, that a game contains (i) any in-game purchases and (ii) loot boxes specifically. An empirical examination of relevant adverts on Meta-owned platforms (i.e., Facebook, Instagram, and Messenger) and TikTok revealed that only about 7 % disclosed loot box presence. The vast majority of social media advertising (93 %) was therefore non-compliant with UK advertising regulations and also EU consumer protection law. In the UK alone, the 93 most viewed TikTok adverts failing to disclose loot box presence were watched 292,641,000 times total or approximately 11 impressions per active user. Many people have therefore been repeatedly exposed to prohibited and socially irresponsible advertising that failed to provide important and mandated information. Implementation deficiencies with ad repositories, which must comply with transparency obligations imposed by the EU Digital Services Act, are also highlighted, e.g., not disclosing the beneficiary. How data access empowered by law can and should be used by researchers is practically demonstrated. Policymakers should consider enabling more such opportunities for the public benefit
Swarm-inspired controllers: a comparative study of decentralized behaviors for distributed manipulation surfaces
No full textThis paper proposes a self-organizing decentralized control framework for modular robotic manipulation surfaces composed of locally interacting actuators. These surfaces aim to induce object translation and rotation using only local sensing and actuation, without centralized coordination or global object tracking. We introduce and systematically compare four behavior rule variants–Discrete, Logistic, Gaussian, and Fourier–under the umbrella of Swarm-Inspired Controllers. Through simulation experiments on various 2D object shapes, we evaluate positioning accuracy, orientation alignment, operation time, and robustness to actuator failure. Results show that decentralized behavior achieves positioning, orientation, and fault tolerance comparable to a centralized baseline, despite relying solely on local information. This demonstrates that fully decentralized heuristics can match centralized control in effectiveness, while offering scalability and resilience
Neural Cellular Automata for Decentralized Sensing Using a Soft Inductive Sensor Array for Distributed Manipulator Systems
No full textIn Distributed Manipulator Systems (DMS), decentralization is a highly desirable property as it promotes robustness and facilitates scalability by distributing computational burden and eliminating singular points of failure. However, current DMS typically utilize a centralized approach to sensing, such as single-camera computer vision systems. This centralization poses a risk to system reliability and offers a significant limiting factor to system size. In this work, we introduce a decentralized approach for sensing in Distributed Manipulator Systems using Neural Cellular Automata (NCA). Demonstrating decentralized sensing in a hardware implementation, we present a novel inductive sensor board designed for distributed sensing and evaluate its ability to estimate global object properties, such as the geometric center, through local interactions and computations. Experiments demonstrate that NCA-based sensing networks accurately estimate object position at 0.24 times the inter-sensor distance. They maintain resilience under sensor faults and noise and scale seamlessly across varying network sizes. These findings underscore the potential of local, decentralized computations to enable scalable, fault-tolerant, and noise-resilient object property estimation in DMS
Tracing Human-AI Relations: A Participatory Approach to GenAI Integration in Creative Public Service Work
No full textThis paper examines a participatory process for integrating Generative AI (GenAI) into the creative work of a public service organization. Through this process, we gain insights into the complexities of human-AI relations, the evolving nature of creative work, and the conditions necessary for meaningful AI integration in service organizations. The paper makes two key contributions: it documents a case of participatory GenAI integration in a public service setting, and it explores implications for creative practice and the evolving role of human-AI collaboration. Findings highlight the importance of reflection, value-driven AI integration, and the need for facilitated spaces for critical reflection. The study contributes to ongoing discussions on how service organizations and designers can engage with AI in ways that align with professional identity, ethics, and creative agency
1st Workshop on Data Craft as Boundary AI Practice
No full textData is central to AI performance, yet its curation remains an invisible and labour-intensive process, often leading to biases and reliability issues. While HCI has explored methods to improve data work, a growing body of research embraces data imperfections in an artistic vein to provoke reflection on human-AI relations. This workshop examines "Data Craft"- practices that creatively manipulate data to challenge conventional AI narratives and lower barriers to public engagement. By framing data craft as a boundary practice, we explore its potential to foster dialogue on AI capabilities, limitations, and societal impact. The workshop will investigate how data craft can be systematically integrated into participatory AI efforts, moving beyond artistic spaces to inform public debate. Through hands-on exploration, we aim to uncover strategies for leveraging data craft to engage diverse communities in shaping AI discourse
Exploring the Entanglement of Relational Design, Spatiality and Places
No full textOver the last two decades, we have witnessed a relational turn in design research, embracing the entanglements of human and non-human elements in socio-material arrangements. Despite the foregrounding of relations in these entanglements, little attention has been given to what role spatiality and places may have in relational design. This workshop explores the entanglement of relational design, spatiality and places based on design case studies from participants. Using a design charette methodology, participants will collaboratively (1) map how relational entanglements shape and are shaped by design research and practice; (2) identify and unpack cross-disciplinary tensions and opportunities; and (3) co-create a vision and action plan for bridging silos and fostering collaboration across disciplines. The aim is to build an interdisciplinary understanding of the intimate entanglement of human and non-human elements, spatiality and places in the complex social-material arrangements we engage with in relational design
Danish Game Industry Timeline
No full textThis timeline collects landmarks and events to present an overview of Danish digital game development. It showcases the formation of Danish game development from early grassroot networks and state inquiries to, among other things, the Law of NIMBI GameLab - Denmark's Institute for Game Development coming into effect in 2025
