A hospital-based lab surveyed 1,440 teens on AI in 2024

The design

In November 2024, the Digital Wellness Lab at Boston Children's Hospital released, via its own Pulse Survey landing page, 'Optimism and Uncertainty: How Teens View and Use Artificial Intelligence,' a report by researchers including David Bickham, Nicole Powell and Michael Rich. The report's own methodology section states that a nationwide sample of 1,440 adolescents aged 13 to 17 was surveyed between 2 and 6 August 2024 using the Alchemer online research platform, with follow-up questions limited to technologies each respondent said they had used or heard of.

What the evidence says

The report's own numbers show 773 respondents who had used generative AI tools and 1,121 who had used voice assistants, with 741 answering questions about both; roughly half of the sample, the report states, indicated 'an openness to forming parasocial relationships, or one-way and non-reciprocal emotional attachments' with voice assistants or generative AI tools, and teens more often described voice assistants than generative AI as potential social companions. The report frames this as a self-reported attitude captured in a single-week online survey, not a measured behavior observed over time, and separately reports that about a third of respondents described feeling 'freaked out' by generative AI, a finding the authors present alongside the openness figure rather than as a contradiction to resolve.

What it asks of people

The report explicitly declines to resolve why teens form these attachments or whether the pattern is healthy, stating in its own conclusion that it will be 'necessary to understand why teens are drawn to these types of interactions and relationships, whether they are healthy and fulfilling,' and what safeguards should exist; this is the authors naming an open question, not answering it. It asks researchers, parents and platforms to treat the finding as a starting point for further study rather than a settled account of adolescent chatbot use.

Privacy and safeguards

The published report does not describe the survey platform's own data-retention practices for respondent answers; its methodology section covers sampling, fielding dates and the online panel used, not what data Alchemer or the panel provider retained about the 1,440 teen respondents. The authors, clinical researchers at a pediatric-hospital-affiliated lab, name self-report bias and the cross-sectional, single-week design as boundaries on what the survey can support, without extending the finding into any clinical claim about individual respondents.

• Does a later citation of 'half of teens' specify that this is openness, not a measured relationship, and toward voice assistants and generative AI combined?
• What follow-up, longitudinal survey, if any, has tested whether this stated openness predicts later chatbot use?
• How does the one-third reporting feeling 'freaked out' relate to the roughly half reporting openness — the same or different respondents?

The Pulse Survey's own design, a one-week, self-report, cross-sectional online panel, is stated plainly in its methodology section, and that design supports a correlational snapshot of teen attitudes in August 2024, not a causal or longitudinal account of how AI companionship affects adolescents over time.