New Framework Evaluates Contextual Appropriateness of Empathy in AI
A recent research article presents an economic viewpoint on empathy within artificial intelligence, utilizing signaling theory to analyze interactions between humans and AI. The authors introduce Signal Cost Proxies—emotional richness, perspective-taking, and contextual tailoring—aligned with affective, cognitive, and associative empathy. This comprehensive framework facilitates a structured assessment of empathy, considering not only its existence but also its suitability in relation to user expectations. The study highlights a significant issue: too much empathy may come off as manipulative, whereas too little may be perceived as dismissive.
Key facts
- Paper titled 'Appropriateness of Empathy in AI: A Signal-Cost Perspective'
- Published on arXiv under Computer Science > Human-Computer Interaction
- Introduces signaling theory to human-AI conversations
- Proposes Signal Cost Proxies: emotional richness, perspective-taking, and contextual tailoring
- Maps proxies to affective, cognitive, and associative empathy
- Framework evaluates empathy appropriateness relative to user demand
- Addresses risks of excessive or insufficient empathy in AI
- Submission ID: arXiv:2605.31340
Entities
Institutions
- arXiv