Engineering Cross-Compatible Personalization for Interactive AI-Driven Services

Negotiable Personalization

The workshop challenges current practice of user modelling and profiling in AI and ML-driven services, where sovereignty over user profiles resides with service providers. We focus on the concept of user-sovereign personalization, where individuals can make their digital self-representations editable, inspectable, and selectively shareable across platforms. This topic is especially timely and critical because current AI-driven services often force users into a frustrating trade-off between privacy and platform lock-in, where they must remain loyal to a specific platform to benefit from an accurate user model. Furthermore, the need to resolve this tension is sharpened by emerging regulatory requirements surrounding transparency and user oversight, such as the GDPR and the EU AI Act. To tackle these challenges, the workshop invites position papers spanning multiple computer science and interdisciplinary domains, including human-computer interaction (HCI), distributed systems, artificial intelligence and machine learning, interactive systems engineering, and security and privacy. Specific themes for these contributions encompass developing schemas and modular representations for user profiles, designing interaction techniques for inspecting or editing user models, creating negotiation workflows and consent mechanisms, engineering user-side AI architectures like federated learning, ensuring verifiability and accountability, and establishing empirical methods to measure user agency and predictability.

Invited Speakers

Organizers