Towards Human(s)-in-the-Loop

Abstract

In this invited talk I depart from the history of human-computer interaction and address current research challenges and approaches for making human-information interaction more social in Human(s)-in-the-Loop concepts. A number of seminal papers have looked at the field of Human-Computer Interaction (HCI) from a broad and historic perspective. Grudin's paper on 'Historical Continuity of Interface Design' which he published in 1990 and is a part of HCI history in itself characterises the evolution of user interface research with a focus on hardware in the 1950s, interface as software in the 1960s and 1970s, interface as terminal from the 1970s to the 1990s, and interface as work setting for groups of users from the 1990s onwards [11]. Bannon some twenty years later was 'Reimagining HCI: Towards a More Human-Centred Perspective' and points out that: 'The ambit of HCI has expanded enormously since the field's emergence in the early 1980s. Computing has changed significantly; mobile and ubiquitous communication networks span the globe, and technology has been integrated into all aspects of our daily lives.' [1, p. 50]. Bodker got a lot of attention with the notion of the 'Third-Wave HCI' on which she writes: 'In the third wave, the use contexts and application types broadened, and intermixed? Technology spread from the workplace to our homes and everyday lives and culture. Research in the third wave challenged the values related to technology in the second wave (e.g., efficiency) and embraced experience and meaning-making?'. [4, pp. 24-26]. In front of those developments-some of which can be seen as evolutions with gradual changes, some of which can be seen as more fundamental shifts, and some of which complement each other and keep on existing in parallel-there have always been great examples of concepts and tools to support human-information interaction. In fact, history has shown great combinations of HCI and hypertext. Engelbart's NLS half a century ago was ground-breaking and has been fascinating scholars until today [5]. On top of the great technological contributions such as structured documents and linked text and network communication and the mouse as input device, Engelbart and his team layed the ground for research on the more general theme of augmenting human intellect by providing users with powerful yet easy-to-use tools for working with information and for social interaction amongst users. For instance, the NLS had a audio and video conferencing tool integrated in the text editor. Despite the fact that both the theoretical contributions on the history of HCI since the 1950s and the stimulating tools by Engelbart and others put a strong focus on the single-user and the communication and cooperation among users, the latter social aspects have been partly neglected. In fact, today there is great research on Computer-Supported Cooperative Work and Social Media-the Springer Journal on Collaborative Computing, the ACM CSCW conference, the ECSCW conference, the ACM Group conference are outlets where great progress is constantly documented. Also in our everyday life today we have more tools for cooperation than ever before-from social media tools such as Facebook and Instagram to communication tools such as WhatsApp and Telegram to cooperation tools such as Google Documents or Etherpad. However, the Web and many other information systems are still lonely places for single users. In other words: on the one side the domain of human-information interaction such as on the Web is still predominantly seen as an arena where single users use the system in a solitary manner, and on the other side the domain of cooperative systems in general is supporting social interaction among users but lack support for systematically dealing with information. In my own research I have tried over the years to bridge the gap between information systems and human communication and communication. Some 25 years ago we aimed at bringing human communication and cooperation to the Web [9, 10]. The goal was and has been to bring users together. So, rather than isolating users from each other, to provide them with mutual information on each other (in the literature on Computer-Supported Cooperative Work often referred to as group awareness) as a basis for easy communication, cooperation, and coordination [6]. Such awareness support has the potential to vastly facilitate social interaction, but also faces several difficulties which can be seen as stimulating research challenges (e.g., mutual awareness information and availability need to be balanced, mutual awareness information and each user's privacy need to be respected). We have also applied our concepts for group awareness and social interaction to recommender systems. Here the goal was to develop concepts for group recommender systems GRS) where groups of users receive suggestions for products or services (e.g., restaurants that fit all group members to have dinner together, or movie to watch in the cinema together). We had developed a technological platform for the GRS that scaled up to generate life recommendations for groups on movies based on real-time geo-located movies and preferences of the users [8]. We conceptualised a process model that allowed the group receiving their recommendations to discuss and assess the recommendations in a fair group process in order to arrive at a final choice that maximally satisfies all group members [2, 3]. Finally, we developed an approach for Human(s)-in-the-Loop where the GRS did not only provide suggestions and facilitated a fair group discussion process, but also monitored the progress of the group discussion and mediates the group discussion by pro-actively pushing new suggestions when appropriate [7].