COURSE 26L1100% FREE
Verified 2026-08-14

Google PAIR guidebook patterns

Google **People + AI Research (PAIR)** publishes the **People + AI Guidebook**—methods, best practices, and examples for designing with AI—plus **design patterns**, **case studies**, deep-dive chapters, and a workshop kit. Citation on si...

What it is

Google People + AI Research (PAIR) publishes the People + AI Guidebook—methods, best practices, and examples for designing with AI—plus design patterns, case studies, deep-dive chapters, and a workshop kit. Citation on site: published May 8, 2019; updated May 18, 2021; CC BY-NC-SA 4.0.

<!-- IMAGE: pattern cards: value, expectations, control, errors, trust -->

HIGH PRIORITYCARD UI
◷ IN PRODUCTION

Visual Spec & Architecture Diagram

PAIR patterns as card deck (3–6 example cards): e.g. explainability, feedback, mental models, trust calibration—titles matching chapter themes. Each card: problem / pattern / example UI thumbnail.

Educational Focus: Patterns stick as cards better than a wall of text.

Why it matters

PAIR is the bridge from academic HCI to day-to-day product questions (“Should we use AI at all?” “How do we onboard?” “How do we show confidence?”). Patterns are action-oriented; case studies show real teams applying them (Photos, Flights, Read Along, Tune, BenchSci).

How it works (plain)

Pattern themes (from the patterns index):

  • Determine if AI adds value (vs heuristics)
  • Set expectations; explain benefits not tech
  • Onboard carefully; calibrate trust
  • Balance automation vs user control; supervise; hand back control on failure
  • Explain thoughtfully; decide whether to show confidence
  • Support users when wrong; be accountable
  • Dataset responsibility & privacy transparency

Use guiding questions on the Guidebook home to filter patterns to your stage.

Everyday example

Flight price predictions that explain “prices often rise” in human terms—without a neural-net lecture (PAIR Flights case study theme).

Try it

Open the patterns page. Pick three patterns for your feature. For each, write one “aim for” and one “avoid” line in your product’s language.

Myths

⚠️ Myth: Patterns are decorative UX stickers after the model ships.
✓ Reality: “Is AI valuable?” can kill a project early—and should.
⚠️ Myth: Case studies are ads.
✓ Reality: They show tradeoffs (control vs automation, confidence UX, feedback).

Sources