Relational Archetypes: A Comparative Analysis of AV-Human and Agent-Human Interactions

Amin Oueslati; Antoni Lorente; Robin Staes-Polet

arxiv: 2604.22564 · v1 · submitted 2026-04-24 · 💻 cs.CY

Relational Archetypes: A Comparative Analysis of AV-Human and Agent-Human Interactions

Antoni Lorente , Amin Oueslati , Robin Staes-Polet This is my paper

Pith reviewed 2026-05-08 09:36 UTC · model grok-4.3

classification 💻 cs.CY

keywords AI agentsautonomous vehiclesrelational archetypeshuman-computer interactiontaxonomymixed traffic flowssocietal impacts

0 comments

The pith

Analogies from autonomous vehicle traffic interactions can generate a taxonomy of relational archetypes for human-AI agent encounters.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper reviews how autonomous vehicles modulate traffic in mixed flows with human drivers and draws parallels to the emerging modes of interaction between people and AI agents. It proposes a preliminary taxonomy of relational archetypes grounded in existing Human-Computer Interaction and AV-human studies. The authors argue that this cross-domain comparison can raise fresh questions about the broader societal effects of AI agents while connecting two research communities that share traits of autonomy, rapid adoption, and economic potential.

Core claim

By extrapolating from the problem of traffic modulation by autonomous vehicles in mixed flows, the paper proposes a preliminary taxonomy of relational archetypes for human-AI agent interactions, built on literature from HCI and AV-human studies, to strengthen bridges between the communities and invite debate on agent impacts.

What carries the argument

The preliminary taxonomy of relational archetypes, constructed by mapping AV-human traffic dynamics onto human-AI agent interactions.

If this is right

The taxonomy can generate new questions about the effects of human-agent interactions across societal dimensions.
It strengthens connections between AV research and AI agent research communities.
It can spark scholarly debate on the different types of societal impact that agents may produce.
It invites researchers to expand comparative analyses between AI agents and other autonomous systems.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Policy makers could adapt AV regulatory approaches to anticipate agent-driven economic and social changes.
Designers of AI agents might use the archetypes to anticipate friction points similar to those seen in mixed traffic.
Long-term studies could test whether the taxonomy predicts measurable shifts in human decision-making when agents become widespread.

Load-bearing premise

The patterns observed in how humans and autonomous vehicles share roads transfer closely enough to human interactions with AI agents to support a useful taxonomy.

What would settle it

Empirical cases in which observed human-AI agent behaviors fall outside the proposed archetypes or in which AV traffic modulation patterns fail to predict agent interaction outcomes would undermine the taxonomy.

read the original abstract

Over the last couple of years, AI Agents have gained significant traction due to substantial progress in the capabilities of underlying General Purpose AI (GPAI) models, enhanced scaffolding techniques, and the promise to drive societal transformation. Companies, researchers, and policy makers have started to consider the different effects that AI agents may have across different dimensions of our lives. However, the literature exploring the broader effects of human-agent interactions is still underdeveloped. In this paper, we review the problem of traffic modulation by autonomous vehicles (AVs) in mixed traffic flows and extrapolate the learnings to the different modes of interaction between humans and AVs to the pair humans-AI agents. In doing so, we propose a preliminary taxonomy of relational archetypes based on literature on Human-Computer Interaction (HCI) and AV-human interaction and tentatively explore how the resulting framework may lead to new questions regarding human-agent interactions. Our effort is aimed at strengthening existing bridges between these two research communities, which share similar traits: autonomy, fast adoption, high impact, and great potential for economic transformation. Building on previous analogies between AI Agents and AVs (e.g., regarding autonomy levels), we anticipate this paper to spark scholarly debate on the different types of impact that agents may have on our societies, while inviting other researchers to expand the scope of their comparative analysis regarding AI Agents.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper sketches a taxonomy by analogizing AV traffic to AI agents but leaves the key transferability question open.

read the letter

This paper reviews AV-human coordination in mixed traffic flows and uses that to propose a preliminary taxonomy of relational archetypes for human-AI agent interactions. It draws on HCI and AV literature plus existing autonomy-level analogies, then tentatively flags new questions the framing might raise for agent research. The goal is to connect the two communities around shared traits like autonomy and rapid adoption. That synthesis is the main output, and it stays explicit about being exploratory rather than definitive. The bridging intent comes across cleanly without overreach. The soft spot is the central analogy itself. Physical traffic modulation involves real-time safety, observable intent, and embodiment constraints that differ sharply from digital agents, which are often non-physical, asynchronous, and scalable. The paper does not lay out boundary conditions or test cases where the archetypes would break, so the taxonomy rests on logical extension rather than demonstrated fit. No new data or formal mapping appears. Readers working on conceptual framing in HCI or AI policy could get some value from the structure as a discussion starter. Empirical or mechanism-focused researchers will find little to use directly. It engages the cited literature in a straightforward way. I would send it for peer review so referees can pressure-test the extrapolation and see whether the taxonomy generates useful follow-on work.

Referee Report

1 major / 2 minor

Summary. The paper reviews the problem of traffic modulation by autonomous vehicles (AVs) in mixed traffic flows and extrapolates learnings from human-AV interactions to human-AI agent interactions. Drawing on HCI and AV-human interaction literature, it proposes a preliminary taxonomy of relational archetypes (e.g., involving trust calibration, conflict resolution, and role asymmetry) and tentatively explores implications for new research questions on human-agent interactions, with the goal of bridging the AV and AI agent research communities.

Significance. If the analogy between physical AV traffic dynamics and digital AI agent interactions holds with sufficient fidelity, the taxonomy could serve as a generative conceptual bridge, prompting cross-disciplinary questions on autonomy, adoption, and societal impact. The paper's explicit building on prior autonomy-level analogies between AVs and agents is a constructive strength for community linkage. As an exploratory, literature-based proposal without new empirical data, formal derivations, or validation tests, its significance is primarily heuristic rather than conclusive.

major comments (1)

[Abstract and taxonomy proposal section] The central extrapolation from AV-human traffic modulation (physical constraints, real-time safety, behavioral adaptation) to human-AI agent interactions (typically digital, scalable, non-physical) is load-bearing for the taxonomy but lacks an explicit mapping of core relational dimensions or boundary conditions under which the analogy would fail (e.g., differences in embodiment, intent observability, and latency). This appears in the abstract's description of the extrapolation and the section proposing the taxonomy.

minor comments (2)

[Abstract and Introduction] The abstract and introduction could more clearly delineate the scope of the 'preliminary taxonomy' versus existing HCI frameworks to avoid potential overlap with prior work on human-AI relations.
[Discussion of implications] Clarify whether the relational archetypes are presented as falsifiable predictions or purely descriptive; the current tentative exploration leaves this ambiguous for readers seeking to extend the framework.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive feedback on our exploratory paper. We address the major comment below and will revise the manuscript to incorporate an explicit discussion of the analogy's boundaries.

read point-by-point responses

Referee: [Abstract and taxonomy proposal section] The central extrapolation from AV-human traffic modulation (physical constraints, real-time safety, behavioral adaptation) to human-AI agent interactions (typically digital, scalable, non-physical) is load-bearing for the taxonomy but lacks an explicit mapping of core relational dimensions or boundary conditions under which the analogy would fail (e.g., differences in embodiment, intent observability, and latency). This appears in the abstract's description of the extrapolation and the section proposing the taxonomy.

Authors: We agree that the paper would benefit from greater explicitness regarding the scope and limitations of the AV-to-agent analogy, particularly since the taxonomy is load-bearing. While the manuscript is positioned as preliminary and heuristic (drawing on prior autonomy-level analogies and aiming to spark cross-community debate rather than assert equivalence), we will revise the taxonomy proposal section to add a dedicated subsection on 'Boundary Conditions and Relational Dimensions.' This will map key differences including physical vs. digital embodiment, direct vs. inferred intent observability, and real-time safety-critical latency vs. asynchronous digital interactions, along with conditions under which the analogy may break down. We will also update the abstract to reference these considerations. This revision strengthens the presentation without altering the exploratory nature of the work. revision: yes

Circularity Check

0 steps flagged

No significant circularity: taxonomy rests on external literature review and analogy

full rationale

The paper contains no equations, parameter fits, or derivations. Its central move is a literature review of AV-human traffic interactions (drawn from cited external sources) followed by an explicit extrapolation via analogy to human-AI agent interactions, culminating in a preliminary taxonomy also grounded in cited HCI and AV literature. No step reduces a claim to a self-citation chain, a fitted input renamed as prediction, or a definitional equivalence; the work is self-contained against external benchmarks and does not invoke uniqueness theorems or ansatzes from the authors' prior work.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The claim depends on the domain assumption that AV traffic interactions transfer meaningfully to AI agents, plus the invented construct of relational archetypes introduced without independent empirical grounding.

axioms (1)

domain assumption AI agents and AVs share sufficient traits in autonomy, adoption speed, and societal impact to permit extrapolation of interaction patterns
Invoked in the abstract to justify the comparative analysis and taxonomy proposal.

invented entities (1)

Relational archetypes no independent evidence
purpose: To categorize and organize modes of human-AI agent interactions
New taxonomy proposed as the main output, with no falsifiable predictions or external validation provided in the abstract.

pith-pipeline@v0.9.0 · 5546 in / 1215 out tokens · 52372 ms · 2026-05-08T09:36:47.648242+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

5 extracted references · 5 canonical work pages

[1]

Anderljung, M.; Barnhart, J.; Korinek, A.; Leung, J.; O'Keefe, C.; Whittlestone, J.; Avin, S.; Brundage, M.; Bullock, J.; Cass-Beggs, D.; Chang, B; Collins, T.; Fist, T.; Hadfield, G.; Hayes, A.; Ho, L.; Hooker, S.; Horvitz, E.; Kolt, N.; Schuett, J.; Shavit, Y .; Sid- darth, D.; Trager, R.; adn Wolf, K. (2023). Frontier AI regulation: Managing emerging r...

work page doi:10.1145/3593013.3594033 2023
[2]

Huang, C.; Wen, X.; He, D.; and Jian, S. (2022). Sharing the Road: How Human Drivers Interact with Autonomous Vehicles on Highways. Proceedings of the Human Factors and Ergonom- ics Society Annual Meeting, 66(1), 1437–1441. https://doi.org/ 10.1177/1071181322661165 Jansen, P., Coté, M.-A., Khot, T., Bransom, E., Mishra, B. D., Majumder, B. P., Tafjord, O....

work page doi:10.1177/1071181322661165 2022
[3]

Compute Trends Across Three Eras of Ma- chine Learning,

AI Agents That Matter No. arXiv:2407.01502. arXiv. http://arxiv.org/abs/2407.01502 Kasirzadeh, A.; and Gabriel, I. (2025). Characterizing AI Agents for Alignment and Governance. https://arxiv.org/abs/2504.21848 Lee, J. W.; Wang, H.; Jang, K.; Hayat, A.; Bunting, M.; Alanqary, A.; Barbour, W.; Fu, Z.; Gong, X.; Gunter, G.; Hornstein, S.; Kreidieh, A. R.; L...

work page doi:10.48550/arxiv.2402.17043 2025
[4]

Shi, J.; Jain, R.; Doh, H.; Suzuki, R.; anf Ramani, K

Practices for Governing Agentic AI Sys- tems. Shi, J.; Jain, R.; Doh, H.; Suzuki, R.; anf Ramani, K. (2023). An HCI-centric survey and taxonomy of human-generative-AI inter- actions. arXiv preprint arXiv:2310.07127 Smakman, J.; Soder, L.; Dunlop, C.; Pan, W.; and Swaroop, S. (2025). An Autonomy-Based Classification: Liability in the Age of AI Agents. Neur...

work page doi:10.51593/20240034 2023
[5]

Wells, P.; and Xenias, D. (2015). From ‘freedom of the open road’ to ‘cocooning’: Understanding resistance to change in personal private automobility. Environmental Innovation and Societal Transitions, 16, 106–119. https://doi.org/10.1016/ j.eist.2015.02.001 Wu, C.; Bayen, A. M.; and Mehta, A. (2018). Stabilizing Traffic with Autonomous Vehicles. 2018 IEE...

work page arXiv 2015

[1] [1]

Anderljung, M.; Barnhart, J.; Korinek, A.; Leung, J.; O'Keefe, C.; Whittlestone, J.; Avin, S.; Brundage, M.; Bullock, J.; Cass-Beggs, D.; Chang, B; Collins, T.; Fist, T.; Hadfield, G.; Hayes, A.; Ho, L.; Hooker, S.; Horvitz, E.; Kolt, N.; Schuett, J.; Shavit, Y .; Sid- darth, D.; Trager, R.; adn Wolf, K. (2023). Frontier AI regulation: Managing emerging r...

work page doi:10.1145/3593013.3594033 2023

[2] [2]

Huang, C.; Wen, X.; He, D.; and Jian, S. (2022). Sharing the Road: How Human Drivers Interact with Autonomous Vehicles on Highways. Proceedings of the Human Factors and Ergonom- ics Society Annual Meeting, 66(1), 1437–1441. https://doi.org/ 10.1177/1071181322661165 Jansen, P., Coté, M.-A., Khot, T., Bransom, E., Mishra, B. D., Majumder, B. P., Tafjord, O....

work page doi:10.1177/1071181322661165 2022

[3] [3]

Compute Trends Across Three Eras of Ma- chine Learning,

AI Agents That Matter No. arXiv:2407.01502. arXiv. http://arxiv.org/abs/2407.01502 Kasirzadeh, A.; and Gabriel, I. (2025). Characterizing AI Agents for Alignment and Governance. https://arxiv.org/abs/2504.21848 Lee, J. W.; Wang, H.; Jang, K.; Hayat, A.; Bunting, M.; Alanqary, A.; Barbour, W.; Fu, Z.; Gong, X.; Gunter, G.; Hornstein, S.; Kreidieh, A. R.; L...

work page doi:10.48550/arxiv.2402.17043 2025

[4] [4]

Shi, J.; Jain, R.; Doh, H.; Suzuki, R.; anf Ramani, K

Practices for Governing Agentic AI Sys- tems. Shi, J.; Jain, R.; Doh, H.; Suzuki, R.; anf Ramani, K. (2023). An HCI-centric survey and taxonomy of human-generative-AI inter- actions. arXiv preprint arXiv:2310.07127 Smakman, J.; Soder, L.; Dunlop, C.; Pan, W.; and Swaroop, S. (2025). An Autonomy-Based Classification: Liability in the Age of AI Agents. Neur...

work page doi:10.51593/20240034 2023

[5] [5]

Wells, P.; and Xenias, D. (2015). From ‘freedom of the open road’ to ‘cocooning’: Understanding resistance to change in personal private automobility. Environmental Innovation and Societal Transitions, 16, 106–119. https://doi.org/10.1016/ j.eist.2015.02.001 Wu, C.; Bayen, A. M.; and Mehta, A. (2018). Stabilizing Traffic with Autonomous Vehicles. 2018 IEE...

work page arXiv 2015