Designing for Friction: Using Strategic/Desirable Inefficiencies to Address the Emerging Privacy Challenges of AI
Authors
Annabelle Larose (Palantir Technologies), Naomi Kadish (Palantir Technologies), and Philipp Wotan Wolf (Palantir Technologies)
Abstract
The paper aims to explore the concept of ‘Desirable Inefficiency’ in privacy engineering, highlighting its significance and limitations in product design, particularly in designing and developing AI and machine learning applications. The paper will focus on the analysis of existing privacy features, such as cookie consent interfaces, and will provide recommendations for implementing desirable inefficiency without hindering usability. Furthermore, it will examine use-cases in data analytics and assess the legislative landscape in Europe and the US, proposing potential changes to promote adoption of desirable inefficiencies in critical digital workflows.