Blog: What if AI says “I don’t know”?
Should we assume AI always has an answer to whatever we ask? Information can be ambiguous or missing, and “I don’t know” (from AI or human) can be the honest answer. AI needs to be trained to recognize uncertainty, assess causes, and effectively support human-AI exploration.
Blog: A 2-way Street of Explanatory AI
Are AI systems inherently explainable, or inexplicable? Different types of systems and uses of AI present the need for different types of explanation capabilities — and different challenges to get there. This post explores considerations that shape explanation in human-AI teaming.
