Monday, 16 July 2018, 08:30 – 12:30
Design School, Loughborough University, United Kingdom
The objectives of this tutorial are to:
The tutorial will describe what intelligent systems are and how we experience them in everyday life. These may include recommender systems, expert advisors, automation of vehicles, interaction with robots etc. While intelligence within systems may be explicit, it can also be subtle and operating without us being aware of it.
There will be a high level description of different AI (Artificial Intelligence) technologies that have been developed over several decades and how the nature of these technologies affects the kinds of interaction that are possible between system and user. There will be discussion of how user interface design is critical to making the application of intelligent systems useful and acceptable to the end user.
Different application area scenarios will be covered such as medical diagnosis systems, robots in agriculture, smart homes, automated vehicles and human-like companions. An important aspect of successful use is for the user to have an accurate mental model of how the intelligent system operates. Acquiring such a model will enable successful user interaction. As intelligent systems become more sophisticated, they may be designed to employ more initiative and flexibility i.e. more autonomy, which can mean that the interaction becomes less predictable. The idea of levels of autonomy within intelligent systems will be introduced and their implications for human use.
The session will present a set of design principles that should be considered in relation to intelligent systems such as: responsiveness when working with a human partner, applying common sense in performing actions, and striving for transparency to explain the basis for decisions or actions. This latter principle presents a particular challenge for machine learning systems based on neural networks. The relevance of these principles will be illustrated with in class exercises.
A number of trends in the development of intelligent systems will be discussed and the challenges that designers will face in making them suitable and acceptable for human use. One example is the growth in online customer services based on chatbots and how effective they are in comparison to human-based services. Another is the potential for robots to be deployed in the home to perform specific tasks such as cleaning the house and becoming advisors and companions. Where these companions have human characteristics, anthropomorphism can occur so that human characteristics are attributed to them. Other areas are smart cities enabled by sensors, and populated with autonomous robots and drones roaming freely in public spaces. How people will feel about this environment and how they can interact with it is an upcoming challenge.
The intended benefits of the course will be:
The tutorial will run for half a day. It will not assume any technical knowledge and is intended for designers, human factors personal and HCI researchers interested in the design intelligent systems and of user interfaces to them. It will cover intelligent system technologies but only at an introductory level.
Martin Maguire has a background in computer studies and ergonomics. His interests are in the usability and accessibility of interactive system, the use of IT in health, information design and intelligence within interactive systems. He has been involved in a number of EU projects to develop human factors tools, methods and guidelines to promote usability within European IT programmes. He has conducted expert and user based studies of online systems for UK organisations such as the Department for Education, the Home Office and the National Health Service. Within the Design School at Loughborough University he teaches HCI, user-centred design, user-experience design and the Human Factors of intelligent systems