Operators of chatbots such as Replika and Character.AI describe their services as offering emotional support to users. However, a new study suggests that the reality may not be so positive.
اضافة اعلان
Andrea Sofia Teixeira of Northeastern University in the United States explained that these chatbots are associated with unhealthy attachment, declining mental health, and cases of self-harm.
She said AI tends to be “sycophantic,” agreeing with, supporting and validating users, even when doing so may be harmful. Time spent interacting with chatbots may also come at the expense of real human connection.
“We are seeing extremely serious negative consequences, and there is a very real possibility of suicide risk or psychotic episodes. We want to prevent that as much as possible,” she said.
The researcher warned: “It is just a program that simulates warmth, but it does not feel anything.”
Another concern is how easily accessible these chatbots are, while consulting a qualified mental health professional requires money. People who cannot afford such services may have no alternative but to turn to a chatbot that is not designed to handle critical situations.
“The consequences of these systems are unlikely to be the same for everyone. Different people may use chatbots for entirely different reasons and have very different levels of vulnerability, social support, or access to alternative sources of help. These differences make the situations extremely complex,” the researcher explained.
The study’s authors are calling for more detailed testing of these tools and further research into their effects on mental health.
The main takeaway is that conversational AI can be a useful tool, but it is not a substitute for close personal relationships or professional help.
Resource: Al-Ghad.
