Log in or create a free Rosenverse account to watch this video.
Log in Create free account100s of community videos are available to free members. Conference talks are generally available to Gold members.
Designing with and for Artificial Intelligence
This video is featured in the Josh's test playlist playlist.
Summary
Rapid advances in Artificial Intelligence and machine learning are transforming the world in many ways. For the product designer or design strategy practitioner this megatrend manifests itself in 2 orthogonal dimensions: AI as a product design material – AI enables solutions that are smarter, faster and can answer questions well beyond human capability alone, but you must deploy them effectively and responsibly to be successful. AI designing the product for you – AI generation of competent oil paintings and music based solely on a set of input requirements has been repeatedly demonstrated in the past decade. Emerging AIs can design entire digital user experiences, code them, and deploy to the cloud with one button click. While AI automation can provide huge benefits in both megatrend dimensions it carries spectacular risk when deployed within life and death systems such as autonomous vehicles and medical products. Concurrently, generative AI for product design carries significant liability risk plus the potential of employment disruption for creative and strategic job careers.
Key Insights
-
•
AI in UX splits into using AI as a design material versus AI augmenting or replacing designers in creative processes.
-
•
Soft AI, which uses structured data and domain rules, is more explainable and suitable for critical applications like genomics than hard AI.
-
•
Trust and perceived credibility in AI-driven medical systems depend heavily on both explainability and interface design quality.
-
•
The genomics AI case analyzes massive, changing DNA variant data impossible for humans alone to process in real time.
-
•
Ben Schneiderman’s classification of AI as super tools or teammates helps frame AI’s role in augmenting human work.
-
•
Clean Software’s AI builds entire UX workflows and code through semantic interaction models, speeding up app development for enterprises.
-
•
Generative AI UX designs face risks like sameness and depend heavily on accurate, high-quality input data to avoid creating useless outputs.
-
•
AI can accelerate UX exploration by generating multiple alternatives quickly, supporting iterative design and decision-making.
-
•
Accessibility and localization best practices can be baked into AI-generated UX code automatically.
-
•
Ethical and regulatory oversight become crucial when AI influences high-risk decisions like clinical diagnoses.
Notable Quotes
"If you don’t trust it, then there’s nothing here."
"The AI is looking through material and that material’s changing every day."
"Visual design quality actually affects perceived trustworthiness."
"You can’t evaluate bias if the AI can’t explain itself."
"Garbage in, garbage out—if the requirements are wrong, the AI will instantly create a useless UX."
"You don’t want to game anybody here. This is persuasion by evidence, not by trickery."
"The marketplace is going to decide if it’s close enough in cost-benefit tradeoff."
"AI-generated UX is not about replacing designers, but removing grunt work to focus on higher-order design."
"Human beings understand graphical user interfaces as composed of objects and actions—this grammar is key to AI design."
"The future was already here. It’s just not evenly distributed."
Or choose a question:
More Videos
"We use the RICE method—reach, impact, confidence divided by effort—to prioritize customer requests in our sprints."
Smitha Papolu Nova Wehman-Brown Melissa Schmidt Adam MenterTheme 3 Discussion
June 4, 2019
"Climate is not operating in geologic time anymore—we face radical, rapid changes unlike past eras."
Victor Lombardi Ted Booth HK Dunston Andrew OtwellBridging Design and Climate Science
February 14, 2024
"We track every hour so the team's time is visible and understood as the collective engine of revenue and rest."
Chelsea MauldinLet's Talk About Money
November 17, 2022
"After swapping our remote meeting solution in early 2020, ESAT went from 69% to 92%, a huge improvement."
Adel Du ToitGet Your CFO To Say: 'Our Strategic Goal is User Obsession'
June 10, 2022
"Land acknowledgments don’t actually change the allocation of power or resources; they work at the mythological level."
Victor UdoewaRadical Participatory Design: Decolonizing Participatory Design Processes
December 10, 2021
"Never put numbers on qualitative research findings; qualitative insights are about richness and why, not how many."
Christian RohrerResearch Operations at Scale
November 7, 2017
"Creating one adopted component can be harder than creating a hundred that never get used."
Nathan CurtisDesign Systems for Us: How Many One-Source(s)-of-Truth Are Enough?
January 17, 2019
"Users want answers to urgent benefit questions but also opportunities to explore other programs they might qualify for."
Alan Williams Rose DeebDesigning essential financial services for those in need
February 10, 2022
"Our effectiveness scale measured audience, purpose, documentation, cadence, effectiveness, participation, and culture to rate meetings."
Briana Thomas Christina RodriguezWhen Design Ops Comes in H.O.T. : A Tale of a Transformed Design Org
September 30, 2021
