---
name: tai-ch006-variance-not-all-differences-are-bugs
description: 'Apply chapter 6 of Testing AI, Variance: Not All Differences Are Bugs, as a workflow for evaluating AI and non-deterministic systems. Use for test planning, eval design, quality review, release evidence, examples, or coaching related to variance: not all differences are bugs.'
---

# Variance: Not All Differences Are Bugs

Skill name: `tai-ch006-variance-not-all-differences-are-bugs`

Based on **Testing AI: Engineering Confidence in AI Systems** by **Jason Arbon**.

## Purpose

Good testing distinguishes harmless variation from variation that changes facts, safety,
reliability, or user trust.

## Use This Workflow

- Identify the AI behavior or release decision being evaluated.
- Define realistic cases, slices, unacceptable outcomes, and evidence needed for confidence.
- Choose measurements that match the risk: rubric scores, samples, intervals, traces, human review, deterministic checks, or production monitors.
- Report uncertainty, severe failures, and decision impact instead of only a pass/fail result.

## Key Guidance

Variance is the spread of behavior. In non-deterministic systems, some spread is expected. The
tester's job is deciding which spread is healthy flexibility and which spread is quality risk.
For example, wording variance may be acceptable in a support answer, but factual variance is
not. Latency variance may be acceptable for a batch report but unacceptable for a real-time
assistant.

## Apply The Approach

Create representative cases, score them with explicit criteria, review severe failures separately, report uncertainty, and connect the evidence to a concrete decision.

## Expert Notes

At expert level, variance should be reported by dimension. Score variance, factual variance,
latency variance, ranking variance, and policy variance are not interchangeable. A single
average can hide the type of instability users will actually feel.