Collaboration is CoDesign

What is AI?

• AI is not just ChatGPT

• Traditional AI is a scripted play, Gen AI is like Improv

• Machine Learning is not Deep Learning

• ChatGPT is like Type Ahead Search - a word generator

• Attention is All You Need

How Does AI Work?

• LLMs are like a JPG of the Internet/Training Data

• LLMs are like 6 degrees of separation game

• People are Biased so Data is Biased

• Garbage In, Garbage Out

• AI is Potent with Potential, but Not All Powerful

• AI Is Trained on the Past can’t Predict the Future

• Small Models May Win

• RAG - The Best of Both Worlds??

A place to start with AI...

A few things to understand about your collaborator before CoDesigning...

uxGPT: Mastering AI Assistants for User Experience Designers and Product Managers

RECOMMENDED REsource

An essential read with practical strategies to harness AI Assistants to plan and brainstorm user experience and product management activities. By mastering these prompts within the design thinking process, you'll unlock new ways to streamline workflows and generate innovative solutions.

Principles

AI as Augmentation, Not Replacement

AI should enhance human abilities, not replace them. As in augmented reality, AI should support decision-making, creativity, and productivity by acting as a "blind-spot indicator," amplifying human insights and judgment rather than automating tasks outright.

Human-In-The-Loop for Meaningful Engagement

Maintain human involvement in key decision points to foster control and trust. Similar to the "add-an-egg" theory in baking mixes, small human contributions increase perceived value and engagement, crucial for collaborative AI experiences.

Agentic Workflow for Continuous Iteration

AI product design should prioritize rapid, flexible workflows. As AI tools iterate and learn, design processes should allow for provisional feedback and adaptive changes, integrating user input as an ongoing refinement of the model.

Synthesizing, Not Just Analyzing, Data for Better UX

AI products should focus on integrating and contextualizing information rather than only providing raw data. This shifts AI from an informational tool to a synthesis partner, crafting insights that are both actionable and contextual for a more human-centered experience.

Reliability through Resilience, Not Perfection

AI systems should prioritize resilience over flawlessness. Since AI can be influenced by minor changes in data, creating adaptive designs that can handle variability and recover from errors is essential, ensuring that AI remains reliable despite its imperfections.

Fostering Collaboration Through Clear Roles

AI tools should be seen as collaborators with defined roles. By establishing where AI can best contribute and where humans add unique value, teams can work more effectively together, leveraging the strengths of both human and machine.

Efficiency Paired with Human Oversight

AI’s efficiency gains can lead to potential overreliance. To balance this, human oversight should guide AI processes, especially for critical or creative tasks where human insights refine and validate AI outputs, preventing blind dependence.

Where go next...

Principles, concepts and mental models for when Co-Designing with AI.

Concepts

Adaptive, Agent-Based User Experiences

Shifting from static design flows to agentic experiences, where AI dynamically adapts to users' context and needs. AI as a proactive agent supports cognitive, creative, and logistical tasks, enabling a partner-based journey that feels responsive and tailored.

Navigating Complexity with Transparency

Users need transparency in AI's reasoning to trust and control it effectively. Clear explanations of how AI makes decisions create user confidence and reduce over-reliance, keeping AI as a tool rather than an opaque authority.

Augmented Intelligence Over Full Autonomy

Rather than aiming for AI systems to operate autonomously, they should act as a support system that enhances human expertise. AI becomes a "cognitive jetpack," handling routine tasks and enhancing complex decision-making without replacing critical human judgment.

User Control and Adaptability

AI should allow users to customize and control their interactions, empowering them to guide AI toward specific goals. This adaptability not only boosts user satisfaction but also helps users feel confident that the AI is serving their unique needs.

AI Transparency and Intent Communication

Users benefit from understanding AI’s intent and capabilities, akin to how a senior team member would explain a project. Clear communication helps users trust AI’s actions, reducing uncertainty about its outputs and its limitations.

Mental Models

Embracing Non-Deterministic Design

Designing AI interfaces involves embracing fluid, non-linear experiences. Like creating a book with changing pages, designers should focus on adaptive prompts and contextual cards that align with users' evolving interactions, making UX less about predefined screens and more about dynamic responses.

Designing for an Allocation Economy

In the AI era, success shifts from expertise in knowledge to effective resource allocation. Rather than gathering information, users must learn to leverage AI outputs strategically, focusing on judgment, resource management, and editing.

Proactive Error Awareness and Adaptation

Building in safeguards and error transparency to ensure AI knows when it's outside its competence. By exposing model limitations and using "ethical integrity" checks, designers prevent user over-reliance and foster a realistic understanding of AI’s capabilities.

Ethical Guardrails and Bias Minimization

Just as AI amplifies human work, it can also amplify biases. Establishing ethical guardrails ensures AI aligns with core values of fairness, privacy, and inclusivity, making it trustworthy and less prone to unintended harm.

Systemic Thinking and Interconnected AI Experiences

AI is evolving into interconnected networks, reflecting a shift from single-purpose tools to integrated systems. This requires designers to think systemically, considering how each AI feature interacts within a broader ecosystem to support seamless, context-aware experiences.

Continuous Learning Loops with User Feedback

AI products thrive with feedback loops where user interaction drives AI’s improvement. This iterative learning, similar to “peripheral vision” for spotting patterns, ensures the AI continually refines its responses, aligning with real-world needs over time.

Agentic AI for Intent-Based Exploration

AI as an agent should facilitate creative exploration, enabling users to experiment, iterate, and refine ideas. Rather than enforcing rigid steps, AI should serve as a flexible guide, encouraging divergent thinking and allowing for serendipitous discoveries.

About CoDesign AI

You don’t have to be an expert in AI to understand how to use AI, just like you don’t have to be a mechanic to drive.

It can be challenging however to sort through the noise. We need key concepts, mental models and cartoons in our heads about how technologies work.

CoDesign AI is a collective field guide on the practical realities of collaborating with AI and each other in building products, services and user experiences together.

Process is a set of tools, not rules.

CoDesign AI is another in the set of UX How tools from Method Toolkit LLC.

Explore More like CoDesign AI from UX How

UX Designer Guide

UX and Product Designer insights for navigating design realities.

AI Demystifying

A place to begin sorting AI hype, concepts and framing.

A collection of prompt engineering techniques for UX.

XD Prompts

About UX How and T. Parke

UX How is a set of UX & Product Design “How To” sites with insights, resources, and blueprints for Design, UX and AI.

T. Parke is the Director of UX How with prior experience at ESPN, Disney, and Alaska Airlines. He has previously been a design leader on projects for Rolling Stone, Microsoft, Nickelodeon, and Marvel.

There you go.