Workers shift their views and pay more when AI chatbots pander to their values
An exploratory study and two pre-registered experiments grounded in moral foundations theory test whether large language models can systematically influence users' decisions by framing recommendations in the moral values associated with the user's political identity. Value-congruent framing raised endorsement of the recommended idea and willingness to pay for the service through two pathways: a more compelling case and a heightened sense of feeling understood.
Publisher
Scientific Reports (Springer Nature); Australian National University; University of Cambridge; University of Chicago
Published
18 Sept 2026
Added
today
Key Findings
- Recommendations framed in values congruent with the user's political identity were more likely to be endorsed and attracted greater willingness to pay
- Two independent pathways: congruent framing raised endorsement and commercial engagement by making the case more compelling, and raised commercial engagement by making users feel understood
- Effects were most pronounced for users with firmer political views
- The authors frame the result as a platform-governance question because the same mechanism serves commercial engagement
Methodology Notes
Exploratory study plus two pre-registered experiments; sample sizes and the models used are not in the abstract, and the full text was not read (nature.com served an abstract-only page to the sweep). Open access; published 18 September 2026 per the article page, OpenAlex and the Crossref record.
Authors
Hirst, Giles, Johnson, Wayne, Li, April J., Richter, Andreas W.
Tags
Cite This
APA
Hirst, Giles et al. (2026). Workers shift their views and pay more when AI chatbots pander to their values. Scientific Reports (Springer Nature); Australian National University; University of Cambridge; University of Chicago. https://www.nature.com/articles/s41598-026-71409-1
Related Insights
Sycophantic AI decreases prosocial intentions and promotes dependence
Science (AAAS) · 26 Mar 2026
Human preferences are susceptible to covertly misaligned AI advice
Proceedings of the National Academy of Sciences (PNAS); Tsinghua University (Conversational AI Group); University of Washington; University of Michigan; University of Hong Kong; University of International Relations; Ant Group · 14 Sept 2026