The social power of AI - Algorithmic imaginaries & social media
Recording from AI Lund lunch seminar 23 March 2023
Title: The social power of AI: Algorithmic imaginaries & social media
Speaker: Daniel Møller Ølgaard
When: 23 March at 12.00-13.15
Spoken language: English
In the last few years, our social and political lives have become increasingly interwoven with Artificial Intelligence (AI) technologies. But, at the same time, the influence AI technologies have over us is poorly understood. So how can we, as social scientists generate knowledge about the social power of AI? In providing a preliminary response to this important question, I suggest that we begin by examining what Taina Bucher calls ‘algorithmic imaginaries’—collective ways of thinking about what AI systems are and what they do—as generative of social relationships of power.
I begin by defining and discussing ‘algorithmic imaginaries’ as a valuable conceptual and methodological device with which to take stock of the social power of AI technologies. Next, based on interviews with social media editors across humanitarian organisations, I demonstrate how specific ways of thinking about a particular form of AI technology—the machine learning algorithms that sort and distribute content on social media platforms such as Facebook—generate specific ways of framing humanitarian disasters and discuss the ethico-political implications of this in terms of the relationships between Western spectators and vulnerable others forged hereby.
In lieu of a conclusion, I end by emphasising the value of sustained scholarly attention to algorithmic imaginaries as a way to generate knowledge about the often unintended and sometimes even harmful effects of AI in our social and political lives.