Inauthentic Value Shifts
More than Manipulation
Stefan Buijsman (TU Delft - Ethics & Philosophy of Technology)
Sarah E. Carter (TU Delft - Web Information Systems)
Juan Pablo Bermúdez (University of Southampton)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
In a recent commentary, Aboodi (2025) has criticized our (Buijsman et al., 2025) concern with inauthentic value shifts [IVS] that can occur through human-AI interactions. We presented emerging evidence that such interactions can lead to unperceived changes in values, which can lead to an IVS in one’s practical identity. Such an alienation from one’s identity is, in our opinion, a problem that needs to be accounted for in the design of AI systems. Since AI systems such as LLMs tend to be closely aligned to WEIRD values (see Atari et al., 2023 for empirical evidence that LLMs align with US cultural values), we additionally positioned this as a special concern for non-WEIRD cultures. Aboodi (2025) argues that (1) while inauthentic value shifts may be an issue if they are due to manipulation, they are not an independent problem and (2) the very concern with inauthenticity and autonomy is unique to WEIRD cultures. [...]