Google AI

Design patterns for AI explainability on 3.6B Android devices

Repeatable design patterns for explainability, control and rights-enhancing principles that fed into Google’s People and AI research library.

A gif of two people’s hands holding paper prototypes of mobile phones. On each mobile is a shape that changes as the two hands bring the mockups closer together.
“I was impressed by IF’s work, the speed at which they understood the problem space, and their ideas on the wider system necessary to create trust. IF helped us to understand a complex area for design.”

[.quote-author]Principal Designer, Research & Machine Intelligence, Google AI[.quote-author]


Designing for trust in a near-future world where personal AIs are commonplace. Our work looked at how explainability, control and rights-enhancing principles could practically work in a user interface. This work is confidential so we are unable to describe the work in detail.

A gif of one person’s hands holding a paper prototype of a mobile phone. On the mobile is a shape that changes as the person’s finger tracks backwards on the mockup.
Retrace steps in a local model’s training log. This lets them identify why a model changed, or restore to a previous version.

Our approach

We worked with Matt Jones and Rob Marchant at Google AI, uncovering ways in which people and AI could earn and maintain trustworthy relationships. This work has continued in several other projects with teams within Google AI and Google Research that explored themes of trust, AI and personalisation.

Full stack design and development

Prototypes that connected the emerging AI technologies to what might be possible in the interface.

Design research

Using prompts and provocations in formative research sessions with people, sometimes in their home, sometimes in a lab environment, sometimes online. This helped us to critique our assumptions, build empathy and explore the edges of the problem space. Sarah spoke about some of those edges in a talk at CogX.

Capabilities for trust

Putting words and narratives together that were compelling and useful to internal teams, as unpacked by our client Tiona (~28:00).


Design patterns from these projects have been used across the Pixel market of 3.6B users, and influenced Google’s PAIR library.

A gif of one person’s hands holding a paper prototype of a mobile phone. On the mobile is a shape that changes as the person’s finger tracks a circular shape on the mockup.
A third party acts on the behalf of many users, defining the limits to which different models can learn.

Heading 1

Heading 2

Heading 3

Heading 4

Heading 5
Heading 6

Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat.

Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur.

Block quote

Ordered list

  1. Item 1
  2. Item 2
  3. Item 3

Unordered list

  • Item A
  • Item B
  • Item C

Text link

Bold text




This is some text inside of a div block.

Get in touch

Looking to reduce risk in a product? Have an idea you want to realise?

Book a call  →