select important things

Jane Frances Dunlop

UK

For select important things (2022), I have created my own knowledge system to navigate the shifting landscape of meaning and meaning-making in the 21st century. This knowledge system, a collection of video fragments that combine stock footage with original spoken texts, contrasts my deeply subjective understanding of the world with images of its generalities. Through a series of installations for online and physical sites, the project considers increasingly fraught questions about how humans make knowledge and how that knowledge, in turn, (re)makes the world.

Each iteration of select important things configures and reconfigures the knowledge system videos, organising them with systems and intentions that are obscured from the audience. The videos create ever-changing narratives as they interweave personal representations with those provided by search algorithms and free video libraries. Here, the videos are presented first as a full set and then in a series of arrangements that begin the process of coding and classifying the videos into a knowledge system that is both arbitrary and intentional. 

select important things is a response to two crises of knowledge that have come to define the last decade: the rise of ‘alternative facts’ that undermine and destabilise expertise, and the increasingly complex AI systems for which datasets stand in as knowledge. My aim is to situate these ‘knowledge crises’ within a history of human representation, highlighting the subjectivity of the knowledge systems that play out in fake news or are the basis of machine learning.

The work’s structure is inspired by Knowledge Representation, the field of machine learning that studies how to represent human knowledge for Artificial Intelligence, as well as by feminist approaches to knowledge and contemporary crises of truth. It was sparked by a foundational essay on Knowledge Representation produced by the MIT AI Lab in the mid-90s that considers how to “capture and represent the richness of the natural world” for the systems that teach machines how to use, interpret and create knowledge. select important things probes how ideas become reality, considering how we make the knowledge that structures our world.

It is a reminder that a machine can only learn how and what it is told to, that rumours can become facts through their repetition, that knowledge is a thing we make. Ideas are actions, and actions have consequences.

Jane Frances Dunlop is a Guest Contributor for Panorama.