Explainability scenarios: towards scenario-based XAI design
summary
Introduces explainability scenarios as a concrete design artifact and applies it to an AI activity‑monitoring system for aging‑in‑place that trains on an 18‑month pilot and now watches 35% of clients. The deployed model predicts level‑of‑care transitions from months 1–2 to month 3 at 90% accuracy, but limited pilot size and partial coverage create bias and integration challenges; the scenario shows which stakeholders need which explanations to make operational decisions.
AI-written summary · about
chapters (7)
- 0:00Introduction to XAI design paper
- 2:58Scenario-based design explained
- 4:24Case study in healthy aging
- 5:32Scenario of care level prediction
- 9:27Key explainability questions raised
- 10:30Role of explainability scenarios
- 11:13Q&A and closing
related talks
24mOpen and Understandable: Designing Explainable AI for Everyone
27mAI is Eating the World, But Without AX, It’s Just Code
33mBehind the AI Curtain - Designing for Trust in Machine Learning Products
06mInterview
16mTrust by Design: Building Reliable AI for Sensitive Environments
01h 20mARTOUNTABILITY: ACCOUNTABILITY, AI, AND ART
30mUsing AI to Create Better Requirements and Designs
06mYour AI Product Will Fail Unless You Can Explain It