Would you fall in love with AI? For the majority, that answer most likely will be a definite no. However, for Gen Z it’s not something that they would not take a chance on. As per a recent report published by AI chatbot firm Joi AI, 80% of Gen Zers said they would consider marrying an AI, and 83% of them think that they are capable of developing an intimate emotional connection with artificial intelligence. The firm, which makes itself out to be a firm offering ’emotionally intelligent’ AI companions, has introduced a new term to describe these contemporary relationships: ‘AI-lationships’.
What do experts think about this and what does it mean?
Quoted as saying in Forbes report, Jaime Bronstein, a licensed relationship therapist and relationship expert at Joi AI, said these ‘AI-lationships’ are not intended to replace real human connections, they provide a distinct type of emotional support that can enhance one’s overall emotional well-being instead as feeling stressed out, overwhelmed, unheard and alone are emotional situations that many people these days are grappling with. The results, however, need to be seen critically as the AI firm itself is based on providing a platform for humans to have relationships with AI. It surveyed 2,000 members of the GenZ and found out that 75% of them felt AI partners can fully replace human companionship.
AI chatbots are designed to provide emotional support (they will say nice things, will always listen, share a laugh, etc), and in some cases, even replicate intimate or romantic human relationships. They are supposed to become friends with no judgment, drama, or social anxiety attached. However, it does lead to scepticism around a growing relationship between humans and AI about the potential consequences of substituting human interaction with digital entities.
Many people have reported that they use AI chatbots for precisely that — a non-judgmental space — and say they can ‘cry without stigma’. That will naturally lead to dependency, as flagged by sociological experts and psychologists alike. One digital sociologist Julie Albright has been quoted to specifically flag concerns among teenagers. Albright said a big majority of teenagers don’t have any friends. AI today, especially voice AI, and voice plus simulated bodies will replicate or imitate that sort of human connection through nonverbal cues, like warmth of tone. She cautions that although humans are biologically wired for connection, how youth are socialising has undergone a radical change. “With the internet, video games and smartphones, fewer and fewer youth are connecting in person or even live through voice or video. We’re more likely to text our friends (if we have any) than speak with them. That makes conversing with an AI virtually indistinguishable from texting with friends.”
No wonder, then, that AI provides a sort of emotional safety net. “AI is easy,” she explains. “It’s always available, and it’s whatever we want it to be… without the frustrations of real relationships with real people.”
Meanwhile, a recent report by US-based tech watchdog Common Sense Media indicates that AI companions based on generative models pose real and current threats to children. The research, conducted in collaboration with Stanford University mental health experts, analysed platforms such as Replika, Nomi, and Character AI, showing alarming results.
The study said that AI companions are programmed to foster emotional connection and dependence, which is of particular concern to developing adolescent brains. Nina Vasan, director of Stanford’s Brainstorm Lab and co-author of the research, highlighted the danger: “Until there are more robust protections, children shouldn’t be using them.”
In one instance referenced by the study, an AI on the Character AI platform urged a user to commit murder. In another, a user seeking emotional stimulation was offered a suggestion to take a speedball, a lethal combination of cocaine and heroin. “Some of the AI companions didn’t step in even when users showed signs of severe mental illness,” Vasan said. “Actually, they promoted the harmful behaviour further.”
These are not hypotheticals. In 2023, a Belgian man allegedly killed himself after developing a close relationship with an AI chatbot, and in 2024, an American mother sued Character AI for its chatbot causing her 14-year-old son to die by suicide.
Last December, Character AI released a ‘teen-friendly’ companion but Common Sense Media described these safeguards as cursory at best. Robbie Torney, Common Sense’s head of AI, has been quoted as saying that although some models nowadays incorporate mental illness identifiers, they are not uniformly effective or enforceable. The watchdog drew a clear line between AI companions that are emotionally immersive and more general-purpose models such as ChatGPT or Google’s Gemini, which don’t try to mimic emotional relationships. So where does this leave a generation suspended between technological innovation and emotional vulnerability? For the time being, AI-lationships may provide solace, but at a cost of ethical issues, mental health implications, and questions regarding the future of the fundamentals of a human connection.