Relational Archetypes: A Comparative Analysis of AV-Human and Agent-Human Interactions
Researchers use autonomous vehicle research to classify AI agent relationships...
Researchers Antoni Lorente, Amin Oueslati, and Robin Staes-Polet have published a paper titled 'Relational Archetypes: A Comparative Analysis of AV-Human and Agent-Human Interactions' on arXiv. The work, accepted at the FAST workshop at AAAI 2026, draws a novel parallel between how humans interact with autonomous vehicles (AVs) and how they interact with AI agents. The authors note that while AI agents have gained significant traction due to advances in General Purpose AI (GPAI) models and scaffolding techniques, the literature on broader human-agent interaction effects remains underdeveloped.
The paper proposes a preliminary taxonomy of 'relational archetypes' based on existing Human-Computer Interaction (HCI) and AV-human interaction research. By extrapolating learnings from traffic modulation in mixed flows to the human-agent domain, the authors aim to spark scholarly debate on the societal impacts of agents. They highlight shared traits between the two fields—autonomy, fast adoption, high impact, and economic transformation potential—and invite further comparative analysis to expand the framework.
- Paper uses AV-human interaction research to classify AI agent relationships
- Proposes a taxonomy of 'relational archetypes' based on HCI and AV literature
- Presented at the FAST workshop at AAAI 2026, aiming to bridge two research communities
Why It Matters
Provides a foundational framework for understanding how AI agents shape human relationships, guiding future research and policy.