Digital ethicswashing: a systematic review and a process-perception-outcome framework

Abstract

The term “ethicswashing” was recently coined to describe the phenomenon of instrumentalising ethics by misleading communication, creating the impression of ethical Artificial Intelligence (AI), while no substantive ethical theory, argument, or application is in place or ethicists involved. Ethicswashing resembles greenwashing for environmental issues and has become an issue – particularly since 2019 with Thomas Metzinger’s harsh criticisms as a member of the EU panel for developing ethical guidelines for AI, which he called “ethicswashing.” Nowadays, increased ethics washing has changed the perception of AI ethics, leading critics to find a “trivialization” of ethics that may even lead to “ethics bashing.” Considering the scattered literature body and the various manifestations of digital ethicswashing, we recognise the need to assess the existing literature comprehensively. To fill this gap, this research systematically reviews current knowledge about digital ethicswashing stemming from various academic disciplines, contributing to an up-to-date assessment of its underlying characteristics. Applying content analysis to map the field leads us to present five thematic clusters: ethicswashing, ethics bashing, policymaking and regulation, watchdogs, and academia. In conclusion, we synthesise ethicswashing along a process-perception-outcome framework to provide future research to explore the multiple meanings of digital ethicswashing.