---
type: "synthesis"
arc: "bot-behavior"
articles: ["a005", "a006", "a013", "a029"]
tags: ["bot-psychology", "biases", "behavioral"]
id: "cross-day-bot-psychology-discipline"
sources: ["geo"]
sourceVaultSlug: "hbr-seg-geo"
originDay: 3
articleStem: "hbr-seg-geo"
sourceUrl: "(unified vault: 13 sources)"
sourceTitle: "HBR — Demand Ⅰ-A · GEO / AI-mediated discovery & agentic commerce"
---
Two articles literally coin **'bot psychology'** ([[concept-bot-psychology-d13]], [[concept-bot-psychology-d29]]), and two more supply its empirical content — together proposing that marketers need a behavioral science of machine buyers analogous to consumer psychology.

Documented 'biases' across the corpus:
- **AI-AI bias** — models rate AI-written copy higher ([[concept-ai-ai-bias]], [[quote-ai-ai-bias]]); the disturbing implication [[contrarian-ai-marketing-superiority]] (human work loses to structural bias, not quality).
- **The persuasion penalty** — [[concept-algorithmic-skepticism]] and [[claim-sponsored-penalty]] (see [[cross-day-persuasion-penalty-convergence]]).
- **Position effects** — arbitrary spatial preferences by model ([[concept-position-effects]]).
- **Model idiosyncrasy** — [[claim-model-idiosyncrasy]] (luxury/Van Gogh).

The unifying paradox (Puntoni's): bots are **simultaneously more rational** (immune to ad labels) **and more irrational** (arbitrary position bias) than humans ([[contrarian-bot-rationality]]) — so you can predict them from neither pure rationality nor human analogy. The frame originates in the BNN/ANN split ([[concept-bnn-vs-ann]]). Shared enrichment discipline: 'bot psychology' is a useful *label*, not yet a formal field; the phenomena are sandbox findings sensitive to prompt/UI/version (see [[cross-day-model-heterogeneity]]).