Is it good or bad for humans to form intimate relationships with machines? This question has vexed machine designers for many years. One of its early appearances in computing was Joe Weizenbaum's Eliza program in 1966. Eliza mimicked a conversation one might have with a Rogerian psychotherapist. Weizenbaum was astounded when some of his friends, including his secretary, started divulging personal secrets to the machine and having warm feeling for the machine. He tried to dissuade them by showing them the inner workings of Eliza: a short program with no intelligence, just a short algorithm substituting keywords into user-typed strings. He was unsuccessful. They did not want to be dissuaded. This question has come back into public view with the arrival of large language models, which engage in competent, fluid conversations. It is now possible for robots to have natural language conversations with people. One of the areas where this is happening is companion robots, which have been introduced into long term care homes to provide companionship with residents and alert caretakers when someone has an emergency.

Ubiquity is pleased to present a debate on companion robots. Computer scientist and author Michael Quinn argues their use may bring harmful consequences. Ubiquity's Jeff Riley, a semi-retired technologist and casual researcher in theoretical astrophysics, argues they have proved beneficial in research studies. Following their position papers are short rebuttals by Quinn and Riley on each other's positions.---Peter J. Denning, Editor in Chief, Ubiquity