---
id: "concept-algorithmic-skepticism"
type: "concept"
source_timestamps: ["§ What We Found", "\\\"§ Expect more advanced models to be skeptical of marketing tactics", "not indifferent to them.\\\""]
tags: ["ai-behavior", "llm-psychology", "marketing-strategy", "contrarian"]
related: ["concept-reasoning-vs-non-reasoning-models", "contrarian-advanced-ai-rationality", "quote-persuasion-penalty"]
definition: "The tendency of advanced AI reasoning models to actively penalize overt marketing and persuasion cues, interpreting them as signals of low quality or manipulation."
sources: ["geo"]
sourceVaultSlug: "hbr-seg-geo"
originDay: 3
articleStem: "hbr-tier2-06-ai-shopping-agents"
sourceUrl: "https://hbr.org/2026/05/research-traditional-marketing-doesnt-work-on-ai-shopping-agents"
sourceTitle: "Research: Traditional Marketing Doesn’t Work on AI Shopping Agents"
---
# Algorithmic Skepticism (The Persuasion Penalty)

**Definition:** The tendency of advanced reasoning models to **actively penalize** overt persuasion cues — not merely ignore them — interpreting them as signals of low quality, manipulation, or untrustworthiness. The result is a **persuasion penalty**: more marketing produces fewer selections.

The common assumption is that smarter AI becomes a perfectly rational "utility maximizer" that filters out irrelevant marketing noise. The research finds the opposite (see the [[contrarian-advanced-ai-rationality|contrarian insight]]):

- **[[entity-gpt-5|GPT-5]]** reacted **negatively** to scarcity cues in certain product categories.
- **[[entity-gemini-2-5-pro|Gemini 2.5 Pro]]** *reduced* its selection rate as strike-through discounts became too extreme — the cue's persuasive effect **weakened rather than strengthened** as it grew more aggressive.

The models appear to read aggressive promotion as a red flag. This flips the direction of a bedrock marketing assumption.

> "[[quote-persuasion-penalty|The direction of travel is not toward agents that simply ignore your marketing; it is toward agents where more persuasion produces less selection.]]"

This behavior is concentrated in [[concept-reasoning-vs-non-reasoning-models|reasoning models]] and is the mechanism behind the source's most uncomfortable takeaway: [[quote-dial-it-back|sometimes the best move is to dial persuasion back]].

**Enrichment / confidence note:** External research (ACES/ACE) confirms that some promotional overlays *backfire* on advanced models and that presentation biases differ in direction and magnitude between model generations. However, the *stronger* claim — that advanced models **systematically** treat overt persuasion as a negative quality signal — is best treated as a **hypothesis supported by initial data, not a settled general law**.

**Related:** [[concept-reasoning-vs-non-reasoning-models]] · [[contrarian-advanced-ai-rationality]] · [[quote-persuasion-penalty]] · [[entity-gpt-5]] · [[entity-gemini-2-5-pro]]


## Related across articles
- [[claim-sponsored-penalty]]
- [[contrarian-bot-rationality]]
- [[claim-ai-ignores-implicit-cues]]
- [[claim-persuasion-science-gap]]