Liverpool Hope University technology reporter Stephanie Power admits to feeling uneasy about the new generation of talking AI companions. She recounts an incident where she overheard her husband conversing with a friendly woman’s voice in his home office. It turns out he was using an app called Pi.ai, which is an example of conversational AI. Unlike traditional AI systems that simply answer questions or perform tasks, conversational AI aims to provide a more natural and interactive conversation experience. The app allows users to type their half of the conversation while the AI responds in one of six different human-like voices. The global market for conversational AI is expected to reach $30bn in the next five years. Tech companies such as Google, Microsoft, and Amazon are all developing their own conversational AI companions. The development of human-like voices for conversational AI involves recording actual humans and then adapting the recordings to convey the desired tone and volume. However, developers may avoid strong regional accents as studies suggest that people are less likely to believe what is being said if the voice is difficult to understand. While conversational AI has potential benefits, such as providing companionship for the elderly or offering support in call centres, there are concerns that people may become overly reliant on AI companions and therapists, mistaking them for real people.