Design for Cognitive Bias
summary
Design choices change behavior: anonymizing resumes (Philadelphia prints and redacts, plus a Chrome plugin whose code is on GitHub) cuts gender/race signals; an AI trained on 10 years of biased hires reproduces sexism; an anchoring phone-number experiment shifts bids; Amazon Go removes intrusive staff and reduces profiling; plain language, pictograms and rhyme (Click It or Ticket) raise seatbelt use and roughly save ~4,000 lives; a one-day red-team review, an assumption audit (five questions), and attention to cognitive fluency expose hidden harms before launch.
AI-written summary · about
chapters (26)
- 0:00Introduction
- 0:25Talk overview
- 1:09Cognitive bias basics
- 2:35What is bias
- 3:37Confirmation bias dangers
- 5:46Choice architecture impact
- 6:49Bias in hiring
- 9:19Cognitive fluency
- 11:03Rhyming and truth
- 12:47Clear language in health
- 14:12Rhyming in safety
- 14:54Name pronunciation effects
- 15:56Framing effect
- 18:22Science vs confirmation bias
- 19:05Red team blue team
- 20:50Outside perspective
- 21:52Survivorship bias
- 23:57Speculative design
- 25:03Assumption audit
- 26:47Power mapping
- 29:14Defining jobs
- 31:04Ethical design
- 32:26Tarot of tech
- 32:46Collective action
- 33:50Person-oriented society
- 34:21Q&A
related talks
38mDesign for Cognitive Bias: Using Mental Shortcuts for Good
12mHow we infuse our biases in the products...
47mCreating solutions for everyone
24mEliminating Unconscious Biases in Technology Through...
38mThe Psychology of UX
27mCrouching Harm, Hidden Bias—Addressing issues of fairness....
27mDesigning with a biased mind
20mMicrointeractions for the System: Design in the Age of Techlash