Abstract
Over the years, sensitive data has been growing in software systems. To comply with ethical and legal requirements, the General Data Protection Regulation (GDPR) recommends using pseudonymization and anonymization techniques to ensure appropriate protection and privacy of personal data. Many anonymization techniques have been described in the literature, such as generalization or suppression, but deciding which methods to use in different contexts is not a straightforward task. Furthermore, anonymization poses two major challenges: choosing adequate techniques for a given context and achieving an optimal level of privacy while maintaining the utility of the data for the context within which it is meant to be used. To address these challenges, this paper describes four new design patterns: Generalization, Hierarchical Generalization, Suppress Outliers, and Relocate Outliers, building on existing literature to offer solutions for common anonymization challenges, including avoiding linkage attacks and managing the privacy-utility trade-off. © 2024 Copyright held by the owner/author(s).

Abstract
This research revolves around the potential challenges, opportunities, and strategies associated with human-centered generative artificial intelligence (AI) in the music compositional practice, emphasizing the role of metaphorical design in shaping musicians' expectations toward the adoption of generative AI in their everyday creative activities. Through a human-computer interaction (HCI) lens, this paper aims to discuss the cultural implications of the human-AI metaphorical design space for the seamless integration of intelligent algorithmic experiences in a manner that aligns with cultural values and realistic expectations of music creators while promoting informed policies, sociotechnical imaginaries, and culturally sensitive generative AI design strategies with focus on user-friendly interfaces that resonate with diverse music creation groups.

Abstract

Abstract

Abstract

Abstract