Researchers warn long-term AI use may foster social isolation, call for better data access

An opinion piece by Cornell researchers raises the possibility that extended interaction with chatbots could gradually erode social connections, urging the scientific community to investigate this potential cumulative harm.

The authors note that OpenAI has begun tracking signs of user distress across multiple sessions and has introduced safeguards such as parental controls, usage limits, break reminders, and crisis‑detection referrals to address immediate mental‑health concerns.

Legislators are also stepping in, with several states now requiring chatbot operators to disclose that the tools are not human and to provide crisis‑referral procedures for users.

While these measures target acute risks, the writers argue that the industry and policymakers are overlooking longer‑term effects and call for broader data access so researchers can study whether AI use contributes to isolation over time.