| Research Projects | Promoted | Algorithmic Bias From a User Perspective: Assessment, Impact, Interventions
bidt background

Algorithmic Bias From a User Perspective: Assessment, Impact, Interventions

Texts and pictures created with generative AI can be heavily biased against women and minorities. The aim of this project is to gain insight into the phenomenon, its effects, and potential countermeasures.

Project description

The widespread use of generative AI has created new opportunities but also brought substantive challenges for individuals and societies. Recent research has identified the challenge that texts and pictures created with generative AI can be heavily biased against women and minorities by neglecting them or portraying them in stereotypical roles (algorithmic bias). At the same time, initial research suggests that humans trust AI more than other humans to portray the world in an unbiased way.

Based on the extant literature, including our own research in the fields of generative AI and media stereotypes, we will examine algorithmic bias from a social scientific perspective. The empirical research is structured along three work packages. Our aims are to: a) identify awareness and evaluation of algorithmic bias among different stakeholder groups; b) investigate human responses and psychological effects of algorithmic bias; and c) develop interventions to combat its potential for negative impact.

The project results will be of key interest to scientists worldwide, as well as stakeholders in politics, media and creative industries, education, and the general public.

Project team

Prof. Dr. Markus Appel

Professor of Psychology of Communication and New Media, Julius-Maximilians-Universität Würzburg | Human-Computer-Media Institute