While generative AI promises to democratize creativity, we lack empirical understanding of how creative professionals begin using these systems. Through a 10-week longitudinal study following one artist’s self-directed exploration, we ask three case-bounded questions: what interaction patterns emerged? What dynamics characterized sustainable co-creation for this artist? How might we evaluate successful integration in this case? Our findings from this case study reveal: (1) physical environments and material practices drove digital interaction; (2) specific temporal patterns emerged (short prompt bursts, consolidation periods, multi-day resurfacing cycles) and (3) productive discomfort and sustained tension marked successful sessions. These patterns suggest alternative design spaces worth exploring. We contribute: (1) four empirically-grounded design patterns (park-and-resurface, warm-up modes, affect-to-action bridges, comfort controls) that could be prototyped and tested; and (2) speculative provocations that challenge efficiency-first paradigms. This case study demonstrates that some users may benefit from approaches orthogonal to current design paradigms.

Recommender systems are increasingly employed by journalistic outlets to deliver personalised news, transforming news curation into a reciprocal yet insufficiently defined process influenced by editors, recommender systems, and individual user actions. To understand the tension in this dynamic and users’ preferences and perceptions of their role in personalised news curation, we conducted a study with UK participants aged 16-34. Building on a preliminary survey and interview study, which revealed a strong desire from participants for increased agency in personalisation, we designed an interactive news recommender provotype (provocative design artefact) which probed the role of agency in news curation with participants (n=16). Findings highlighted a behaviour-intention gap, indicating participants desire for agency yet reluctance to intervene actively in personalisation. Our research offers valuable insights into how users perceive their agency in personalised news curation, underscoring the importance for systems to be designed to support individuals becoming active agents in news personalisation.

Applications of Artificial Intelligence (AI) in the domain of Personal Health Informatics (PHI) offer potential avenues for personalised treatment and support for people living with long-term conditions, however, they also present a number of ethical challenges. Whilst participatory approaches can help mitigate concerns by actively involving healthcare professionals, patients, and other stakeholders in design and development, these are constrained by the limits of epistemic standpoints and the risks posed by extrapolation from individuals to groups. In this paper we draw upon interviews with stakeholders involved in Human Immunodeficiency Virus (HIV) care, including clinicians, insurance providers and pharmaceutical industry representatives, to map intentions and ethical considerations for developing PHI tools for people living with HIV. Whilst treatment efficacy for HIV has improved patient quality of life and life expectancy, management and care is complicated by knowledge gaps about what living and ageing with HIV entails. We investigate how the critical concept of epistemic injustice can inform the design of data-driven technologies intended to address these gaps, helping orient expert perspectives within the broader structures and socio-historical influences that shape them. This is of particular importance when designing for marginalized populations such as people with HIV (i.e. who may experience social stigma and be under-resourced, managing multiple conditions), helping to identify and better account for fundamental ethical considerations such as equity.

We present a body of work undertaken in response to the challenge outlined by Harper et al. in their paper, ?'What is a File?' [9]. Through a conceptual and design-led exploration of new file metaphors, we developed the 'file biography', a digital entity that encompasses the provenance of a file and allows the user to keep track of how it propagates. We explored this through prototyping and utilised it in two user studies. In the studies, we (i) asked people to sketch out file biographies for their own content, and (ii) deployed a tool enabling users to build their own simple file biographies across multiple versions of Word documents. We conclude that new file metaphors may need to play different roles for different types of digital content, with a distinction being drawn between content that is 'in production' and virtual possessions that are, in a sense, a 'finished' artefact.

This paper responds to an increasing interest in how data-driven interactive artworks can support a greater understanding of the relationship between humans and data. This paper extends this work by focusing on the The Prediction Machine, an interactive and data-driven artwork that arose from collaborations between an artist, climate scientists, HCI researchers and the wider public. A study of the practice-led process reveals: (a) how the artist "performed" scientific data across multiple outputs; (b) how the artist walked a line between scientific and artistic process; (c) how the public experienced data across multiple components of the artwork; (d) how the artwork can be seen as process rather than a exhibition or artifact; and (e) how tensions between artistic strategies and scientific processes inherent in the creation of data-driven artworks might inform future work involving public engagement with scientific data.

The past four years have seen the rise of conversational agents (CAs) in everyday life. Apple, Microsoft, Amazon, Google and Facebook have all embedded proprietary CAs within their software and, increasingly, conversation is becoming a key mode of human-computer interaction. Whilst we have long been familiar with the notion of computers that speak, the investigative concern within HCI has been upon multimodality rather than dialogue alone, and there is no sense of how such interfaces are used in everyday life. This paper reports the findings of interviews with 14 users of CAs in an effort to understand the current interactional factors affecting everyday use. We find user expectations dramatically out of step with the operation of the systems, particularly in terms of known machine intelligence, system capability and goals. Using Norman's 'gulfs of execution and evaluation' [30] we consider the implications of these findings for the design of future systems.

Notions like 'Big Data' and the 'Internet of Things' turn upon anticipated harvesting of personal data through ubiquitous computing and networked sensing systems. It is largely presumed that understandings of people's everyday interactions will be relatively easy to 'read off' of such data and that this, in turn, poses a privacy threat. An ethnographic study of how people account for sensed data to third parties uncovers serious challenges to such ideas. The study reveals that the legibility of sensor data turns upon various orders of situated reasoning involved in articulating the data and making it accountable. Articulation work is indispensable to personal data sharing and raises real requirements for networked sensing systems premised on the harvesting of personal data.

The regulatory climate is in a process of change. Design, having been implicated for some time, is now explicitly linked to law. This paper recognises the heightened role of designers in the regulation of ambient interactive technologies. Taking account of incumbent legal requirements is difficult. Legal rules are convoluted, uncertain, and not geared towards operationalisable heuristics or development guidelines for system designers. Privacy and data protection are a particular moral, social and legal concern for technologies. This paper seeks to understand how to make emerging European data protection regulation more accessible to our community. Our approach develops and tests a series of data protection ideation cards with teams of designers. We find that, whilst wishing to protect users, regulation is viewed as a compliance issue. Subsequently we argue for the use of instruments, such as our cards, as a means to engage designers in leading a human-centered approach to regulation.

Terms and conditions are central in acquiring user consent by service providers. Such documents are frequently highly complex and unreadable, placing doubts on the validity of so called 'informed consent'. While readability and web accessibility have been major themes for some time in HCI, the core principles have yet to be applied beyond webpage content and are absent from the underpinning terms and conditions. Our concern is that accessible web pages will encourage consent, masking the complexities of the terms of usage.

Ubiquitous computing systems tend to be complex, seamless, data-driven and interactive. Reacting to both context, and users' implicit actions resulting from the lived experience, they cast all traces of human life as potential 'data'. To augment users' endeavours, such systems are necessarily embedded below the line of human attention, drawing upon new and highly sensitive types of data. This begs the question, where is the moment of user consent and how can this moment be truly informed? We would argue that it is time to revisit our design principles in respect of consent and redress the balance of agency towards the user. We draw upon a series of multidisciplinary interviews with experts to (a) reframe consent for ubicomp, and (b) offer three indicative principles, supportive of consent, for designers to 'balance' against system functionality. We hope that this will afford a new prism through which designers might make value judgements.

With its emphasis on ‘smart environments’, the vision of pervasive computing raises critical concerns with respect to consent. When sensors capture data about people, and digital systems interpret and respond to that data below the line of user visibility, two fundamental questions arise. First, are current notions of consent relevant in the emerging class of pervasive systems and, secondly, what are the practical consequences of dealing with consent for such environments? This paper reflects on the key principles of consent and the challenges raised by pervasive systems through a review of multidisciplinary perspectives on consent and technology. The developing complexity and decreasing visibility of pervasive computing systems, coupled with the increasing value and sensitivity of personal data, mean that it is no longer sufficient to design systems that assume users capable of making informed decisions at a single moment. In particular, the unprecedented sensitivity of contextual data, and the potential harms associated with inferences made on the basis of that data, highlights the need to revisit our design principles. Many of these discussions are nuanced and implicate a broad range of perspectives; however, it is clear that there is unlikely to be a ‘moment of consent’ in pervasive systems. In order to progress this agenda we offer the following set of recommendations to designers, as considerations for future systems design: (i) electronic consent mechanisms (ECMs) must cease to be designed around ‘moments in time’ and allow for negotiation, (ii) systems should enable establishment of user expectations and development of norms, (iii) systems should be sensitive to third-party interactions and (iv) we should move beyond designing for user control towards designing for user autonomy.

The developing complexity and decreasing visibility of pervasive computing systems, coupled with increasing value and sensitivity of personal data, mean that it is no longer sufficient to design systems that assume users capable of making informed decisions at a single moment. In particular, the unprecedented sensitivity of contextual data, and the potential harms associated with inferences made on the basis of that data, highlights the need to revisit our design principles in respect of consent. This thesis will use a mixed-methods approach to reframe 'consent' for ubiquitous computing systems, resulting in a series of design guidelines to inform future developments.