9 January 2019
As more interactions are taking place through voice-enabled technologies and chat, AI is increasingly making predictions on behalf of users. All too often, data scientists are left to answer questions and make assumptions about user expectations and needs that have experiential consequences. For example, what percentage accuracy does the AI need to have before users find it valuable? How should an e-commerce site determine which ads to display to users? Left without UX intervention, these kinds of questions go unanswered and AI can easily fall into the Uncanny Valley at best or be disturbing at worst.