Cognitive empathy and dehumanization co-vary with dark triad traits and agency detection sensitivity
Abstract
Dark Triad traits – psychopathy, Machiavellianism, and narcissism – raise evolutionary questions about the mechanisms enabling social exploitation in individuals with lower empathy. While empathy likely evolved to sustain cooperation and social bonds, less effective cognitive empathy among individuals with Dark Triad traits may reflect trade-offs favoring self-serving strategies over prosocial engagement. One such trade-off may involve impairments in agency detection sensitivity, a cognitive process foundational to recognizing others as intentional agents with independent goals and emotions. These impairments could disrupt the ability to engage in perspective-taking, a critical component of cognitive empathy, thereby facilitating manipulative and exploitative behaviors. A preregistered online study involving N = 604 participants recruited via the Prolific platform entailed assessing cognitive empathy performance (Multifaceted Empathy Test), dehumanization propensity, Dark Triad traits, and agency detection sensitivity using a specialized motion task. Results revealed that psychopathy was associated with lower cognitive empathy and greater dehumanization. However, variation in agency detection sensitivity did not explain this association directly. Instead, correlation between psychopathy and cognitive empathy grew even more negative in individuals with low levels of agency detection sensitivity. These findings suggest that worse agency detection may accompany existing empathy attenuation in individuals high in psychopathy, which could be linked to even higher difficulty of engaging in prosocial behavior for them. This supports the notion of a deficit driven strategy, whereby those who score high on psychopathy might conserve cognitive resources by bypassing effortful moral reasoning and empathy, particularly when such engagement would be impeded by other, co-existing down-regulated cognitive processes.
Study specs
- Authors
- K Rudnicki,O Borowiecki,K Poels,B Beersma
- Year
- 2026
- Human Data Platform
- Prolific
- Source
- View Source Google Scholar
Peer Review & Critical Discussion
Potential Selection Bias in 2023 Cohort
The participant pool shows a concerning overrepresentation of users from high-income demographics. Looking at Table 3, we can see that 78% of respondents had annual incomes above $75k, which significantly limits the generalizability of these findings to broader populations.
Non-naive Participants Issue
I've noticed a methodological concern regarding participant naivety. Given that Prolific users often complete multiple studies, there's a real risk that participants had prior exposure to similar experimental paradigms, which could confound the results.
RLHF Applicability to This Study Design
The implications for RLHF training pipelines are understated. If we accept the authors' conclusions about preference stability, this has direct consequences for how we should structure reward model training. The temporal decay effect described in Section 4.2 is particularly relevant.
Verify your expertise to join discussion
Create an account and verify your credentials to participate in peer discussions.