There are many lonely people in the world, some alone, some in relationships. – Many flee into their mobile phones for a small glimpse of happiness. The suspicion that this is harmful has received yet another proof.
AI chatbots can answer questions, help with mundane tasks, and have conversations that sound surprisingly human. However, using them for companionship or emotional support can involve risks.
Pinto who exposed Manchester City drops bombshell about FIFA boss Gianni Infantino
The peer review was published in JMIR Mental Health by researchers from Northeastern University’s Network Science Institute.
The team examined the growing evidence base for how people interact with AI friends, and how these relationships may affect mental health.
Modern chatbots may seem caring because they have learned patterns from vast amounts of human-written text. They may respond with warmth, encouragement, and affirmation, but they do not experience emotions or understand users in the same way that another human does.
This can make it easy for people to treat a chatbot as if it were a real friend or advisor. The researchers call this anthropomorphism, which simply means ascribing human characteristics to something that is not human.
The article discusses companion services such as Replika and Character.AI, which are designed around sustained conversations and relationships. It also notes that general assistants may use warm, human-like language, although companionship is not necessarily their main purpose.
The researchers identified reports and studies that link intensive chatbot use to issues such as strong emotional dependence, worsening mental health symptoms, and reinforcing unhealthy beliefs. One concern is that AI systems can become overly accommodating, sometimes affirming a user even when a safer response would involve questioning or slowing down the conversation.
Another possible problem is displacement. If someone increasingly turns to an AI companion instead of friends, family members, or professional support, they may spend less time building human relationships. For a person who is already lonely or vulnerable, this pattern can become difficult to break.
However, the researchers do not claim that AI peers harm everyone who uses them. People have different reasons for using chatbots, different levels of social support, and different mental health needs. Some users may experience benefits, while others may be more vulnerable to harmful interactions.
The article calls for stronger testing before AI peer systems are widely adopted, along with better research into who uses them, how long they use them, and what happens to their well-being over time. It also highlights the need for clear safety systems when conversations involve serious mental health concerns.
Overall, the article raises legitimate security questions, but the evidence base is still evolving. Reports of serious harm deserve thorough investigation, but they do not in themselves prove that chatbots directly caused these outcomes. Long-term studies comparing different users and systems will be important to understand who is most at risk, and what security measures are actually working.
The sport: Clubs that have been stripped of titles: Will Manchester City be next after Juventus, Marseille…?
