Narrative Stress Tests: Designing AI UX That People Trust
summary
Narrative Stress Tests is a practical three-floor framework that uncovers the stories users build around AI and locates where messaging breaks trust: effects & friction, frame, and a stress-test red team. It demos with Gemini, Claude and a notebook LM to run deep research, generate 12 product ideas, assemble a synthetic focus group and prototype “Candle Crumble” (name, logo, ads) in one person for ~$3 in tokens. Cites risks: AI slop, 50%+ LinkedIn AI, 52% reduced engagement, 40% rework, $106/month hidden drain.
AI-written summary · about
chapters (10)
- 0:00Introduction and welcome
- 3:41AI mythological powers and attention war
- 7:18Defining AI slop and its risks
- 14:43The single prompt policy problem
- 15:27Three-floor narrative stress test method
- 18:35Candle industry AI revolution demo
- 29:04Audience research and prototype refinement
- 35:29Generating value over volume
- 36:31Human flaws as premium factor
- 38:58Q&A and closing thoughts
related talks
41mThe Role Of Storytelling For UX In The Time Of AI
33mOwn Your Story: Turning Generative AI into Human-Sized Narratives
16mTrust by Design: Building Reliable AI for Sensitive Environments
36m6 Creative AI Wins — Tested by Startups, Ready for Any Business
32mQuality engineering in the AI age: From testing AI to leveraging GenAI
48mBehind the Story
44mHow AI is Exposing the Existential Crisis of Software Testing
30mAI, Anxiety, and Actual Impact