Designs trust-calibrated AI interactions and error-management strategies for human-AI collaboration.
The AI-First Design skill provides a systematic framework for designing AI-powered features that appropriately calibrate user trust and manage the consequences of AI errors. By applying established research from Yang et al. and the Google PAIR guidebook, it helps development teams classify interaction patterns based on trust-risk and error-risk levels, design effective feedback loops, and implement progressive disclosure strategies. This ensures that human-AI handoffs are safe, intuitive, and account for the non-deterministic nature of LLM outputs.
主要功能
01Progressive disclosure strategies for graduated AI onboarding
02Interaction pattern selection for human-AI collaboration models
03Trust-risk and error-risk classification for AI outputs
04Feedback loop design for handling AI uncertainty and errors
05AI transparency assessments based on Google PAIR patterns
0616 GitHub stars
使用场景
01Auditing existing AI interfaces for trust calibration issues and transparency gaps
02Determining the appropriate level of human-in-the-loop oversight for autonomous systems
03Designing conversational UX and agentic interfaces for LLM-powered features
