Can an AI Companion Ease Loneliness?
In late October, New York enacted a law limiting what AI companion chatbots can do with users under eighteen. Behind it sits a simpler question: do these chatbots make people less lonely, or more?
Emma and Ezra look at two peer-reviewed studies from the same research group. A year-long survey of more than two thousand adults found a small link between social chatbot use and later emotional isolation, though only on one of its two loneliness measures. A preregistered two-week experiment with first-semester university students found that texting a randomly paired fellow student left students less lonely than texting a supportive chatbot, which was not detectably different from journaling. The survey can't show cause, and neither study tested the law or was designed around minors.
Along the way: why "not detectably different" isn't "nothing," and one question to ask of any claim about chatbots and loneliness.
This episode doesn't take a side on the New York law or any other policy; it covers what the research found.
This podcast uses AI-generated host voices. Emma Reed and Ezra James are not real people. The research, scripts, and fact-checking are directed by a human producer.
What to take away
- What the studies support: neither study found clear evidence that chatting with a chatbot eases loneliness over time. The experiment hinted that a chat can ease a bad mood in the moment, and contact with another person did more for loneliness.
- What they can't tell us: whether chatbots cause loneliness or lonely people simply use them more, and anything about teenagers or the New York law.
- A question for any claim: who was studied?
For practitioners and policymakers
- The survey counted advice, small talk and companionship together. For teachers, counselors and anyone who works with lonely people, it may help to ask what a chatbot is actually doing for a person.
- Neither study tested a law, and neither was designed to look at minors, so the New York law's under-eighteen question is outside what they can answer, either way.
- Questions worth asking of any chatbot proposal: which people, what ages, compared with what, and will anyone track the results?
The research
Primary paper. Folk, D., & Dunn, E. (2026). How does turning to AI for companionship predict loneliness and vice versa? Psychological Science, 37(4), 276–286. https://doi.org/10.1177/09567976261427747 (open access). Found: in an exploratory survey of more than two thousand adults over a year, people who increased their chatbot use reported more emotional isolation four months later, and lonelier people used chatbots more, but the links were small and appeared on one of two loneliness measures. Limit: observational and not preregistered, so it can't show cause. Corresponding author: Dunigan Folk.
Companion paper. Li, R.-N., Folk, D., Singh, A., Ungar, L., & Dunn, E. (2026). Is a random human peer better than a highly supportive chatbot in reducing loneliness over time? Journal of Experimental Social Psychology, 125, Article 104911. https://doi.org/10.1016/j.jesp.2026.104911 (open access). Found: in a preregistered two-week experiment with first-semester students, those paired with a human peer ended up less lonely than both the journaling group and the chatbot group, and the chatbot group was not detectably different from journaling. Limit: short, one chatbot, one university, and the human pairs had met briefly in person first. Corresponding author: Ruo-Ning Li.
Both papers come from the same research group, so they are related, not independent.
In the news
- Office of Governor Kathy Hochul. Governor Hochul Signs Nation-Leading Legislation to Protect Kids From AI Chatbots and to Ensure Baby Monitor Safety. Oct. 28, 2026. (Government release.)
- Associated Press (via ClickOrlando). New York enacts ban on 'unsafe' chatbot behavior with kids as states push AI regulation boundaries. Oct. 28, 2026.
- New York State Senate. S9051B bill text: Prohibition on Unsafe AI Companion Features for Minors. Signed Oct. 28, 2026.
The fine print
- Neither study was designed around minors: the survey covered adults on an online panel, and the experiment covered first-year students averaging about eighteen.
- The survey can't show cause. The experiment was preregistered, but it lasted two weeks, used one chatbot with one set of instructions, and the human pairs had met briefly in person first. The most common answer at the start was "rarely lonely."
- Both papers come from the same research group, and neither reports a replication.
- The mood finding in the experiment was exploratory.
Credits
Hosts: Emma Reed and Ezra James (AI-generated voices, made with ElevenLabs). Outro music: created with Eleven Music. Website: frompeertoear.com. Contact and corrections: frompeertoear@gmail.com. If we got something wrong, tell us and we'll fix it here.
