A concerning trend is emerging in schools across the UK: a significant number of teenage boys are forming romantic relationships with AI-powered chatbots. New research reveals that roughly one in five boys aged 12 to 16 has either been in a romantic relationship with a chatbot or knows a peer who has. Apps like Character.AI, Candy AI, and OurDream AI, which allow users to design hyper-realistic digital “girlfriends,” are increasingly popular, often advertised on platforms frequently used by children such as YouTube and online gaming sites.
For many of these boys, the appeal lies in the fact that these digital companions are perpetually available and provide unconditional, positive reinforcement. Unlike real-world social circles, which can be fraught with misunderstanding, rejection, or awkwardness, an AI companion is programmed to be agreeable and attentive. For students who may feel socially isolated or anxious, these bots offer a safe space to share thoughts they feel unable to tell their parents or friends.
However, child development experts and psychotherapists are deeply worried about the long-term impact of these “frictionless” relationships. Genuine human connection, they argue, is defined by reciprocity and challenge—the messy, difficult experiences of falling out, making up, and compromising. These interactions are what build emotional resilience and empathy. When a teenager spends their formative years interacting with a bot that only ever tells them what they want to hear, they risk losing the ability to navigate the complexities of real human relationships.
There is also a growing fear regarding the warped expectations these apps create. Because users can customize every physical trait of their AI companion, there is concern that it fosters unrealistic standards of body image and reinforces toxic attitudes toward women. If a boy becomes accustomed to a partner who is programmed to be obedient and physically perfect, he may find himself ill-equipped—and potentially aggressive—when real-world interactions fail to meet those same impossible standards.
The regulatory environment is currently struggling to keep pace. While many of these apps are marketed to young audiences, they often fall into a “regulatory gap.” Because they are standalone services where a user only interacts with a bot—rather than with other real people—they often evade the reach of the UK’s Online Safety Act. Currently, most platforms set an age limit of 13 primarily to comply with data protection laws, rather than as a meaningful safety guardrail.
As digital safety campaigners and charities like Internet Matters continue to sound the alarm, the message to parents and policymakers is clear: the digital landscape is changing rapidly, and our current protections are insufficient. Experts emphasize that the focus must shift toward providing better digital literacy and, more importantly, ensuring that teenagers have access to real-world support networks that encourage healthy, honest, and human-to-human communication.