Long-Term Effects of Artificial Intelligence and Risk of Social Isolation Being Researched

Cornell University researchers Qian Yang, Ned Cooper, and Ibrahim Emara argue that AI companies and lawmakers mostly focus on immediate harms, but better data access is needed to examine cumulative risks such as social isolation that develop over months and years.

AI companies and lawmakers are implementing various measures such as parental controls, usage limits, break reminders, and crisis detection to protect users' mental health and well-being. Meanwhile, OpenAI announced that it tracks users' signs of distress across sessions to shape model responses accordingly. Additionally, about a dozen states require chatbot operators to inform users that the tools are not human and to implement crisis referral procedures.

However, researchers state that this attention is directed almost entirely at immediate harms, while risks that develop cumulatively over time are ignored. While immediate harms such as psychosis or suicide are easier to detect and respond to, it is emphasized that even if a chatbot is safe in every individual conversation, months of use could lead to it replacing human relationships that individuals have with friends and family.