Designing Human-AI Experience
summary
Presents six concrete HAX principles and seven anti-patterns and shows real outcomes: an AI-powered banking agent (voice cloned from a real employee) that hands off tasks to sub-agents, supports multilingual voice↔text and demonstrably speeds some flows though users slow down when switching to typing; a mobile car-scan app deployed with a human‑in‑loop for Norwegian rental staff that prevents false charges; cites numbers—about 763 million illiterate adults and typical daily speech ~16,000 vs typing ~3–4,000 words—to argue for button-first hybrids and mitigate latency, hallucinations, and forced automation.
AI-written summary · about
chapters (20)
- 0:00Introduction to talk
- 3:36Why human-AI experience matters
- 7:10Frustrations with current AI interfaces
- 10:45AR headsets and future tech
- 13:22Interface regression to command lines
- 17:22Input modes and user expectations
- 19:56Designing hybrid AI interactions
- 23:18Key problems in AI applications
- 26:53Avatars and cognitive limitations
- 32:16Assistants vs agents in AI
- 34:12Impact of AI commerce on users
- 34:55Human-AI experience design principles
- 37:30Seven deadly sins of AI design
- 44:32Co-creation and user involvement
- 47:19Avoiding AI as a marketing gimmick
- 48:45Neglect and prompt paralysis
- 51:30Hidden features and discoverability
- 54:22Tyranny of forced automation
- 56:31Automation vs augmentation debate
- 57:00Q&A and course details
related talks
55mDon’t Make Me Prompt: Intelligent Interfaces Users Actually Like
20mHuman above AI
36mEmbodied AI: A Cognitive Shift, Not a Cosmetic Change
40mHuman-AI Collaboration Models for Better Decisions and Faster Workflows
28mHuman Computer Interaction: The UX of AI
34mWhen Software Understands the User
49mHow to be an AI First company
42mKeeping It Human in A World of Automated Everything