Lingua Franca
A Design Language for Human-Centered AI

Introduction

Despite all the talk of transformation, anything built with AI is secretly a mess.

Under the hood, you’ll find a cacophony of noisy data, opaque algorithms, and false signals leading to all sorts of awkward and unintended results. And yet, organizations are hastily integrating Artificial Intelligence into all facets of our lives, from shopping to health care, from our workplaces to our legal systems.

AI Needs a Design Language

This document is the sum of our collective imagination about how to build, test, measure, critique, and improve—in short, to design—the AI technology rapidly proliferating around us. In it, we hope you will find a wealth of practical, actionable information, regardless of your education or profession or skill set. This is a design language for AI—a standard set of techniques, frameworks, visuals, messaging, and overall design patterns that apply broadly to different kinds of AI to make it more usable, more trustworthy, and better aligned with people.

AI Needs to Care About Humans

Imagine finding yourself in a hospital where you didn’t know why you were admitted, what procedures you were going to receive, whose judgments were being followed. Now imagine being a doctor in this hospital, not knowing which drugs you were administering, which patients you were treating, what conditions you were testing them for. This is, in a Kafka-esque way, our current digital world where AI drives our clicks and purchases, our dating lives, our eating choices, our newsfeeds, our opinions, and soon, our physical bodies.

Design can bring clarity, intuition, and usability to these kinds of experiences. But more importantly, design is a lens with which we should make change in the world. Not just in designing toys and interfaces, but also in designing algorithms, business strategies, policy frameworks. Designing for children, the elderly, the non-tech literate. Designing for novices and professionals alike.

Lingua Franca is intended to help you learn what AI is truly capable of, to help you shape your next AI product or project, to help you pose the right questions to your team and organization, and to ask the crucial question of all of us—how do you wish to live in a world of pervasive intelligence?

Why the name Lingua Franca?

A ‘Lingua Franca’ is a shared working language that helps people of varied backgrounds communicate. Interestingly, Lingua Franca was the name of an actual pidgin language that arose over a thousand years ago to help traders communicate between cities in Spain, France, Italy, and the Ottoman and Arab empires. We think the field of AI needs a lingua franca of design to help communicate ideas through common metaphors that can span across professions, mindsets, and cultural contexts.
Lingua Franca: Overview of principles, elements, and handbook in the form of a galaxy

Layout

Lingua Franca is a design language that functions as both a toolkit for creators and a manual for public consumption. Such a design language for AI technologies cannot simply include a set of typography and visual styles. AI is governed by complex relationships between data, computation, humans, and communities. As such, our design language is broken up three pieces: a Handbook for designing AI systems, a set of eight core Principles of human-centered AI, and a collection of reusable Elements that form the backbone of our work. We developed Lingua Franca from our experience with all kinds of AI systems, and by seeing the connective tissue between them. Our principles and elements work coherently together, in almost any context we have encountered, spanning domains as diverse as computer vision, natural language, forecasting, conversational UIs, and other interactive AI systems.

Authors

Abhay Agarwal is the founder of Polytopal where he manages intelligence and strategy. He has previously held positions as a research fellow at Microsoft Research, a lecturer at the Stanford d.School where he created the course Designing Machine Learning, and a co-founder of location analytics startup Denizen.

Marcy Regalado (co-author) is a human-centered designer and recent graduate of the Stanford Design Impact Master’s program. She won a 2017 Webby award for her interaction design work for Fidelity.

Acknowledgments

  • We would like to thank the following individuals for their feedback and edits: Chris Donahue, Christopher Dowd, David Estrada, Steve Gershman, Noopur Goel, Barry Katz, Mala Kumar, Jules Sherman, Apoorva Tadepalli, Chris Zannetos