How will patterns and experiences evolve in a world shaped by Artificial Intelligence?

AI is accelerating change. Companies are reacting by “doing AI”without prioritizing the experience of the people on the receiving end–Design is more important than ever. To respond, we must get smart, fast. This starts with understanding the emerging patterns of interaction, affordances, and heuristics in an AI world. The pages below are summaries of the current state. Things are moving fast. Get involved!

AI Interaction patterns

Identifiers

Help users identify and distinguish AI features and content

Color scheme

Visual cues to help users identify AI features of content

Iconography

Images that convey the form the AI takes in a product

Name

How do we refer to the AI?

Wayfinders

Give clues to help users understand how the AI works and how to work with it

Nudges

Alert users to actions they can take to use AI in their existing tools

Suggestions

Solves the blank canvas dilemma with clues for how to prompt

Template

Structured templates that can be filled by the user or pre-filled by the AI

Prompts

Different ways that users can generate results with AI

Auto Fill

Makes it easy for users to extend a prompt to multiple inputs at once

Open chat

Open ended prompt inputs that can be used in conversations or to tune results

Remix / Blend

Combine prompts with each other or other sources to get new results

Synthesis

Distill or reorganize complicated information into simple structure

Tuners

Tools to refine your prompt or the results to get more predictable outputs

Filters / Parameters

Let users define constraints to improve the quality of their results

Model management

Let users specify what model to use for their prompts

Personal voice

Ensure outputs match your voice, tone, and preferences in a consistent way

Trust indicators

Let users control the response and assess its accuracy

Caveats

Inform users about shortcomings or risk in the model or the technology overall

Controls

Manage the flow of information or pause a request mid-stream to adjust the prompt

Feedback

Signal expectation gaps or errors in the model – but is that clear to the user?

Footprints

Give users the ability to trace the relationships between sources and results

Follow along

[↑] Subscribe to the newsletter for free to support the project or follow along.

[↑] Join the Slack where we can discuss emerging patterns, new products, and the changing role of design. The community is free to join and participate in.