Individual Submission Summary
Share...

Direct link:

Other People’s Problems: Launching a Research Program for the Study of Ordinary AI Risk

Sun, August 9, 12:00 to 1:00pm, TBA

Abstract

There is widespread concern that the rise and spread of artificial intelligence (AI) will displace economic opportunities and erode human skills. These tandem social problems are referred to as displacement and deskilling, respectively. Here, we present two studies at the foundations of a burgeoning research program, addressing perceptions of displacement and deskilling as a personal vs public risk. Specifically, we report on surveys distributed to music industry practitioners (Study 1) and undergraduate students (Study 2) to examine how participants assess the relative threat of AI on their own and others’ jobs within the creative industry, and on their own and others’ academic skills within higher education. Across both studies, we hypothesize that participants will perceive their own risks of displacement and deskilling as significantly lower than their colleagues and peers (as relevant). That is, the threats of AI are “other people’s problems.” We tentatively term this phenomenon comparative AI optimism. Hypotheses are based on scoping interviews conducted prior to survey construction, bolstered by theories of optimism bias and fundamental attribution error that together predict and explain discrepancies in self vs. other risk assessments. Data collection is still in process, with results indicating preliminary hypothesis support. We will report on early findings from both studies, showing quantitative trends and qualitative open-ended responses. These findings, grounded in creative industries and academic settings, will set the stage for a broader research agenda that spans domains of possible AI-related displacement and deskilling.

Authors