.png){kind=icon}
.png){kind=icon}
Return
Alignment with Psychological Concept Network
.png){kind=icon}
PRE
AI
DOI：10.1007/978-981-95-4384-7_36.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Current approaches to human-AI alignment often separate values, concepts, and behaviors into distinct domains. However, concepts and representations are inherently network-structured, shaping how values are expressed and actions are executed. Human behavior is also influenced by culture, nationality, and social context. We propose conceptnetwork alignment, an approach that connects representational networks with behavioral outcomes using methods from psychology and neuroscience. This framework highlights the interdependence of concepts and actions within complex cultural and multimodal contexts. We review existing models of human behavior and neural representation, identify opportunities and challenges, and suggest evaluation criteria at behavioral, predictive, representational, and mechanistic levels. We argue that progress in this direction is essential for AI systems that areinterpretable, safe, culturallyadaptable, andhuman-centered.
Keywords:
AI alignment
Psychological Concept Network
Multi-level Alignmen
Journal
N
IF:
0
Papers:
26
Citations:
0