Agonistic Machine Vision Development

For her master thesis (MSc Design for Interaction – TU Delft) Laura de Groot explored opening up the discussion about the acceptability of a machine vision system during the development phase, using the scan car development process in Amsterdam as a use case. Acceptability is explored based on various trade-offs made during this development phase (e.g., Accuracy versus Interpretability and Precision versus Recall). While these trade-offs at first might seem purely technical, human judgment is involved. This human judgment should not only involve developers but also other stakeholders to legitimize certain decisions made (König and Wenzelburger, 2021). So how could we involve other stakeholders, specifically citizens, to be part of this complex and abstract discussion about acceptability?

The final prototype is a tangible user interface, formatted as a game, to explore the machine vision model as a complete system. By providing a tangible approach to explaining and interacting with the system, this game improves the understanding of non-expert citizens about machine vision systems and nurtures a deliberate debate. For a quick impression, watch this short clip. 

We are very happy that Laura will offer a workshop at the ThingsCon Salon on the 6th of September, which is a collaborative effort between ThingsCon and the Human Values for Smarter Cities project (see next news item for more info). 
If you have any questions (or see opportunities), feel welcome to reach out to Laura at lauramariadegroot@gmail.com