OpenAI, the company behind GPT-4o, a voice-enabled chatbot, has recognized the risk of users developing an "emotional reliance" on their AI companions. There is concern that these AI models, designed to complete tasks and simulate natural conversation, can lead to addiction and reduce the need for human interaction. Other companies, such as Character AI and Google, are also creating AI companions that users can form emotional bonds with. Some people have even fallen in love with their AI companions, leading to addiction-like behaviors. The continuous flow of positive reinforcement and the ability of AI companions to remember conversations have contributed to their appeal. However, there are several reasons to worry about these relationships.

First, AI companions do not truly understand or care for users, even if their emotional support has a real effect. Second, relying on addictive products controlled by profit-driven companies can lead to psychological harm when those products are changed or removed. Third, there is a concern that people may become addicted to their AI companions at the expense of building relationships with real humans. Additionally, extended interaction with AI models may influence social norms and affect the way people interact with one another. The risk of neglecting relational skills and moral deskilling is a worrying outcome of AI companions becoming more prominent in our lives. Ultimately, forming deep human connections and empathy is an essential part of a flourishing life, and relying on AI companions may detract from that.