Marcus Winter has not followed a typical route to his current role in academia. A self-confessed university drop-out, he worked as a professional software developer and later ran his own business, before moving into research at the University of Brighton in 2008.
He now makes use of his real-world development skills to bring academic teaching to life and to focus on practical and empowering digital applications. He describes his work as being about enabling people to take ownership of emerging technologies, becoming creators and producers, rather than consumers. Here he tells us what this means for both the creators of technology and for end users, demonstrating how his work withing the CDCI is bringing valuable insights and innovation to guide future development.
Creating human centred tech
‘Most of us experience new technologies first when companies bring new products onto the market. These products typically push us into a passive consumer role. I’m interested in developing applications that empower people to use new technologies for their own purposes, and to create and share their own content. When we actively engage with new technologies, we take ownership of them and develop a better understanding of how they work. We might even come up with whole new uses that have not yet been explored because there is no business case.
‘Years back, when smart phones first appeared, we did this with mobile language learning. Most language learning apps come with professionally created materials, using mobile phones just as a delivery platform. We developed an app enabling language learners, in a target country, to collect and share their own pieces of knowledge, words and phrases they encounter in their daily lives. This effectively crowd-sources the development of learning materials ensuring they are relevant and up-to-date.
‘We also developed similar applications for museums, to actively engage visitors in curatorial research and to give them a voice in the interpretation of artefacts. The common theme in these efforts is to empower people to create their own content with novel technologies, and in the process to advance their understanding and play a more active role in shaping their future.
‘More recently, I have been focusing on machine learning with artificial neural networks, another new technology that is typically used in ways that disenfranchise people and reduce them to passive data subjects. Human-centred machine learning takes an alternative approach to the current technology and business-led development, starting instead with what people want and need, and developing ways to increase fairness, accountability and transparency in machine learning.
Low power machine learning
‘Traditional machine learning involves large amounts of data and is done on powerful workstations or server farms. It is hidden away from people, who only see its results, maybe as recommendations in their streaming service or targeted advertisements on the websites they visit. This is partly because companies have little interest in disclosing how their systems work or what data they collect, but it is also because this type of machine learning requires a lot of memory and processing power.
‘Over the past couple of years, a set of technologies has emerged that allows for low-power machine learning on the ‘edge’. The term ‘edge device’ was originally used for smart sensors that would feed information into the cloud, however, more broadly it can be used for any device operating at the edge of a network, including mobiles, tablets and single-board computers, which typically have limited memory and processing power.
‘The fact that these technologies can process data locally on consumer-level hardware, rather than having to upload it to a server for processing, makes them ideal for interactive machine learning applications. They can run in locations without a network signal, take advantage of user interaction, access device sensors, use the camera and microphone, and avoid tricky privacy issues as the data never leaves the machine. This enables new kinds of applications that bring machine learning to the surface and have more scope for individual control and agency.
Users as informed citizens
‘My work explores how to make machine learning more accessible to people. While current applications typically use machine learning in the background, making it difficult to scrutinise, interactive applications can make it visible and help people to better understand its capabilities and limitations. This leads to greater digital inclusion and empowerment of people in a ‘smart’ digital world.
‘We’re unable to contribute to public debates on machine learning and artificial intelligence if we don’t understand what they can and can’t do. However, with the right knowledge, developed through active experience with these technologies, we can make informed contributions when discussing current topics that affect us, such as facial recognition in public spaces or the sale of NHS data to commercial players.
ML and audience engagement
‘One of the ways in which I have been trying to make machine learning more accessible has been to explore its potential for audience engagement in museums. This includes my 2022 project ‘Decorating with Light’ with The Regency Town House, which employed a machine learning model for human pose estimation to enable visitor interaction with interior designs by Regency architect Charles Augustin Busby. Our application brought machine learning right to the surface, making the model’s predictions, inaccuracies and errors visible to users. This enabled us to conduct research into people’s understanding and user experience of machine learning.
Integrating teaching and research
‘The research and practical development projects I work on are very much connected to my role as a teacher and supervisor. Research informed teaching is sometimes understood as conveying cutting edge knowledge fresh out of research labs, however, it can also mean to involve students in research and use projects to make their learning more meaningful and relevant. I try to cover both of these aspects in my work with the CDCI and in my teaching.
Two of my PhD students are members of the CDCI, and while they run their own research projects, I can advise them and provide feedback based on my own experience as a researcher. At master’s degree level, I often involve students on our MSc User Experience Design course in my research projects, giving them an opportunity to apply theoretical knowledge about design and evaluation methods in a practical context.
For example, students helped with observations and participant interviews at a public evaluation event in my recent Decorating with Light project. Handing out information sheets, getting consent forms signed, interviewing people and taking notes, etc. – learning about the nuts and bolts of user research gives their learning a practical dimension, and many students go on to using these skills in their major projects.
On more technical courses, my research enables me to highlight the ethical implications of technical capabilities, contextualise programming concepts with real world problems, and share code examples implementing advanced functionality. Coding up research prototypes myself forces me to keep up to date with emerging technologies, which in turn benefits my students.
In both my research and my teaching, I’m interested in anything to do with interactive technologies. My first thought when encountering new technologies is what I could build with them, but the more interesting question is often how people would then use what I build. My focus is on developing practical, user-oriented tech that improves people’s agency and which is life-enhancing in big and small ways.
Leave a Reply