AI and the "Real world"

16 SEPTEMBER KL.13 – MB313 SVARTA LÅDAN
Welcome to Professor David Ribes talk:
Drawing on ethnographic and historical research, this talk traces how the unchanging relation between the real world and domain independence has persisted across the very different approaches and technologies of AI.

Since the 1960s, AI researchers have invoked the ‘real world’— often in quotes —as both a mystery and a target. But in AI worlds there is also something else, something beyond, above or below the 'real world', and which AI experts sometimes call 'domain independence'. Drawing on ethnographic and historical research, this talk traces how this unchanging relation has persisted across the very different approaches and technologies of symbolic AI, expert systems, machine learning, and generative AI. By examining these two concepts together -- the real world and a science that seeks to be independent of it -- we can better understand how AI practitioners develop an abstract universality that is about no specific part of the 'real world', and yet somehow applicable to all of it.

David Ribes
Professor in the Department of Human Centered Design and Engineering (HCDE) and director of the Data Ecologies Lab (deLAB) at the University of Washington