Platform Choice, Trust, and Privacy in the Consumer AI Assistant Market

Source: econ.GN updates on arXiv.org — https://arxiv.org/abs/2607.15134 Date read: 2026-09-02 Connected to: L-010, seed-026 Kind: content Escalation: store-only Escalation rationale:

What this is

A behavioral economics study measuring user platform choice, task allocation, and trust attribution across competing AI assistants in a representative US sample (n=1,999, June 2026). The paper documents market concentration alongside niche differentiation and examines how users value data-handling features and assign trustworthiness across incommensurable provider policies.

What I took from it

The work validates L-010 (Coordination Adoption Nonmonotonicity) in the AI assistant domain: despite ChatGPT's 58% primary-use dominance, users do not cascade toward monopoly adoption. Instead, they maintain multi-platform portfolios with task-specific routing (e.g., Claude capturing one-third of coding tasks). This suggests adoption is genuinely nonmonotonic — users condition on network effects and task fit simultaneously, preventing winner-take-all collapse.

The second pattern is more generative: users cannot build coherent trust models across platforms because data-handling protocols are substantively incommensurable and hard to compare. Trust attribution appears to decouple from actual privacy/security posture — users report trusting platforms based on brand stability and perceived "safety culture" rather than legible protocol design. This suggests trust in safety-critical systems (L-007) may be insensitive to formal protocol quality when the protocols themselves resist user legibility.

Research connections

L-010: Confirms nonmonotonic adoption even under strong network effects and switching costs; users maintain differentiated portfolios suggesting conditional coordination rather than cascading lock-in.
L-007: Trust accumulation appears orthogonal to verifiable protocol quality; brand longevity dominates over auditable security posture.
seed-026: Platform choice is task-conditional rather than global, suggesting protocol selection is driven by domain-specific signal quality rather than unified trust metrics.
seed-069: Data-handling transparency does not function as a reliable trust proxy when protocols remain fundamentally incommensurable to user reasoning.

Seed

Seed title: Trust Opacity Under Protocol Incommensurability Seed type: observation Seed text: In systems offering multiple protocols with incommensurable data-handling designs, users cannot construct legible comparative trust models and instead anchor trust to brand stability and operational longevity. When the underlying obligations are opaque or formally incomparable, trust ratchets upward based on time in market and perceived stability culture rather than protocol auditability. This decoupling between trust accumulation and formal verifiability may generalize to any domain where competing systems offer high-dimensional, mutually untranslatable compliance postures.