---
id: "concept-ai-friction"
type: "concept"
source_timestamps: ["§ The Takeaways for Managers", "¶14", "¶15"]
tags: ["metrics", "user-experience", "ai-adoption"]
related: ["concept-hidden-coordination-costs", "action-measure-friction", "contrarian-adoption-vs-friction"]
definition: "The measurable behavioral resistance and extra effort users exhibit when struggling against an AI system, visible in logs via repeated rephrasing, long exchanges, and arguments."
sources: ["tail1"]
sourceVaultSlug: "hbr-seg-tail1"
originDay: 1
articleStem: "hbr-tail-113-ai-personality-problem"
sourceUrl: "https://hbr.org/2026/06/does-your-ai-have-a-personality-problem"
sourceTitle: "Does Your AI Have a Personality Problem?"
---
# AI Friction

AI friction is the **behavioral evidence** of the [[concept-hidden-coordination-costs|hidden coordination costs]]. It is the measurable, extra effort employees spend wrestling with an AI system instead of deriving value from it.

Crucially, **friction can coexist with high adoption**. Employees may log in frequently and generate high query volumes simply because they are *mandated* to use the tool, while quietly struggling against it — the vanity-metric trap examined in [[contrarian-adoption-vs-friction]].

Friction is observable in **standard usage logs without special equipment**. Key indicators:

- Abnormally long back-and-forth exchanges
- Repeated rephrasing of the exact same request
- Rising attempts to argue with, bypass, or override the AI's instructions

The practical response is to [[action-measure-friction|mine ordinary logs for these signals]] rather than trusting satisfaction surveys, which [[claim-self-reports-fail|fail to detect friction at all]].


## Related across articles
- [[concept-operational-noise]]
- [[claim-contextual-performance-variation]]