It’s a pleasure to share this article by Prof Chris Speed who is Chair of Design Informatics at the University of Edinburgh. Over to Chris…
A great deal has changed since the Creative Informatics R&D Cluster began in October 2018. Data-Driven Innovation as a term is becoming common place, and AI has moved from an imaginary based upon the images and tales of science fiction, and in to practical examples that give the public an idea as to their limitations. We now know that Gartner was almost right in 2015 when they predicted that “by 2017, a significant disruptive digital business will be launched that was conceived by a computer algorithm” although it wasn’t a business, it was politics, and we are all familiar with the distortions that algorithms are making to the news feeds that inform our decision making.
As designers we have seen the slow creep of AI tools, and this year all of the big creative application developers have extolled the virtues of their systems from Adobe Sensei to Autodesk’s 3D and VFX Content Creation. Central to the ‘sell’ of these tools is their role in complimenting the creatives skills and not usurping them, such as Autodesk’s aim to produce ‘Trusted Collaborators’ as the third level of their product intelligence (beyond smart tools and intelligent assistants): “Tools that understand the context in which we [the human] work. These systems turn data into contextual insights.” (Evan Atherton, Senior Research Engineer).
Whilst Adobe and Autodesk are keen not to challenge the creative control of the creative, colleagues and I developed the international ThingTank project that pursued the idea that it wouldn’t be long before things would start designing things themselves, and they would no longer require humans to help. The project was funded to explore the potential for identifying novel patterns of use within data that is streamed through the interaction between people and things, and things and things. Through an understanding of what data can tell us about how we use domestic objects in practice, the project posited that new models of use would emerge and reinvigorate the role of things and people within design and manufacturing.
In the past, many Internet of Things projects have used the network connection of artefacts to identify cost saving and process efficiencies (e.g., vehicle manufacturers), or to track goods within large networks (e.g., logistics companies), or to monitor the health and safety of systems (e.g., aircraft manufacturers). Such projects look for regular patterns within datasets which suggest efficiencies that will reinforce the identity of a product or service by making its function easier to use or more economical. By contrast, the ThingTank project proposed that looking for anomalies and outliers in datasets could suggest more radical design opportunities. During studies, the research team developed non-anthropocentric methods by gathering and streaming data from both material objects and humans that were involved in a domestic relationship, to better understand how machines could identify practices that went unidentified by human researchers (Giaccardi et al 2016).
Although the majority of us use products as intended, many of us also invent novel usages of objects by adapting or using them for unintended purposes. By scanning large datasets for evidence of mis-use and then using them to build new assemblages, the ThingTank project proposed that algorithms may exploit data to design things that human designers could have never have conceived.
Our early prototype was written to look for anomalies in the use of domestic forks, and performed image searches across the Google image database to find any use of a fork that it felt wasn’t perhaps normal such as for use in gardening, painting and as a wrist splint. Unfortunately the funding was stopped abruptly due to the oil war in 2016, so we will never know what our algorithm would have designed if it had been allowed to move to the next stage – use the insights from the transgressive use of forks to design new products for humans. Nevertheless two design fictions emerged in response to intelligent things: Uninvited Guest by Superflux that reminded us that humans are as likely as AIs to deploy transgressive tactics to ‘make do and get by’, and Teacher of Algorithms by Simone Rebaudengo that introduces a school for bad behaving data-driven artefacts:
Although ThingTank never reached its full potential it heralded a way of understanding how the use of data is shifting within design. Following the work, I worked with the late Prof. Jon Oberlander to develop the Ablative Framework for Design Informatics to reflect on their existing methods of working with data, in order to anticipate its ability to transform design process as its level of performativity increases:
Design from data: when systems are designed by people, where they are inspired by measurable features of humans, computers, things, and their contexts.
Design with data: when systems are designed by people, where they take into account the flows of data through systems, and the need to sustain and enhance human values.
Design by data: when systems are designed by other systems, largely autonomously, where new products and services can be synthesised via the data-intensive analysis of existing combinations of humans, computers, things, and contexts.
The framework sees design from data as established methods for designers, and design by data as still highly emergent; whilst design with data is the important space of enquiry that requires urgent research to address the full extent of human data interactions as data-driven technology begins to question as to whether the human is at the top of the creative tree.
About Chris Speed
Prof. Chris Speed is Chair of Design Informatics at the University of Edinburgh where he collaborates with a wide variety of partners to explore how design provides methods to adapt, and create products and services within a networked society. Chris is the Director of the Centre for Design Informatics that is home to a combination of researchers working across the fields of interaction design, temporal design, anthropology, software engineering and digital architecture, as well as the PhD, MA/MFA and MSc and Advanced MSc programmes. Chris is also Director of the recently funded Creative Informatics R&D Partnership, one of the nine AHRC funded Creative Industries Clusters.
The points raised in this article were a key focus addressed at the BEYOND conference in Edinburgh, November 2019. Conference themes explored the impact of AI, Machine Learning and Data on the creative industries. The two-day programme offered lively debate, keynote sessions, short-talks, panel discussions and exhibitions showcasing the UK’s excellence in research-driven creative innovation, plus networking.
Main Image credit: A still from ‘ThinkTank Internet of Things research projects’ video – funded by Skoltech