Various forms of voice assistants are available on mobile phones, home automation or in-car systems.. From an AI point of view they require low-level speech recognition followed by some form of natural language processing to interpret what is said. They may employ techniques designed for text chatbots or large language models. They are often helped by operating within a limited environment, for example, the devices and lights within a home. Crucially the human-computer dialogue has to be carefully designed to ensure a robust interaction in the face of uncertainty in mutual understanding.
Used in Chap. 19: page 318