Designing meaningful Human-AI Interactions
summary
Microsoft presents the Microsoft HACKs toolkit with 18 guidelines grouped into four stages for designing human‑AI interaction; examples show Microsoft Word, Google Photos, Bing, Pinterest and Siri using specific patterns like make clear what the system can do and showing confidence levels. Guidelines run on Azure conversational AI use cases, encourage global controls, granular feedback and planning for failure, and do not stop mistakes but supply controls and explainability to let users mitigate harms.
AI-written summary · about
chapters (22)
- 0:00Introduction
- 1:49Need for AI interaction guidelines
- 8:16Microsoft’s 18 design guidelines
- 10:23Make capabilities clear
- 13:11Show confidence level
- 14:56Time services by context
- 16:44Show relevant info
- 17:25Match social norms
- 22:01Invocation and correction
- 24:30Explain decisions
- 25:55Remember interactions
- 27:00Learn from behavior
- 28:47Encourage feedback
- 29:28Consequences of actions
- 30:10Provide global controls
- 31:58Notify about changes
- 33:01Do guidelines prevent mistakes?
- 37:58Applying guidelines in practice
- 39:23Impact on development
- 41:08Team diversity matters
- 42:11Summary
- 43:15Q&A
related talks
36mStrategies for Human-AI Interaction and Collaboration
03mDesigning for humanlike interaction
57mDesigning Human-AI Experience
27mThe Shape of AI : How designers integrate AI into their workflow
12mDesigning AI To Scale Human Thought
28mHuman Computer Interaction: The UX of AI
25mArtificial Intelligence in the Modern Workplace
53mAI for the better