Designing Artificial Intelligence
summary
Shows a two-minute Google Home rollercoaster where an Italian grandmother cycles from delight to confusion and reveals lack of hearing cues, steep syntax needs and no proactiveness. Describes an Innovation Lab study of elders finding poor stickiness and missing proactivity, a Carnegie Mellon prototype that fuses microphone and EMI sensors to infer door openings, and LivePerson survey of 5,000 people who expect handoff to humans. Cites failures (Uber crash, pumpkin misread, Taobao mistranslation), argues for sensors + AI for context and for preferring clearly robotic form factors to avoid the uncanny valley.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 1:50What is AI today
- 3:14AI’s imperfections
- 7:13Defining human intelligence
- 11:33Building AI interactions
- 13:35User interactions with AI
- 18:47UX challenges in AI
- 21:22Elderly and home assistants
- 23:31Sensors and AI proactiveness
- 26:18AI in autonomous vehicles
- 28:04Skeuomorphism in AI
- 30:36New AI device form factors
- 31:41AI limitations and chatbots
- 33:51AI expectations gap
- 36:21AI’s UX frontier
- 37:04Q&A
related talks
57mDesigning Human-AI Experience
26mArtificial Intelligence? - More like Artificial Stupidity!
26mHumanity and AI
20mEmbracing the Voice Revolution
32mA Human Story: AI, the Home, and Personal Computing
34mWhen Software Understands the User
55mDon’t Make Me Prompt: Intelligent Interfaces Users Actually Like
28mHuman Computer Interaction: The UX of AI