A national survey found most teens have tried AI companions

The design

Common Sense Media's Talk, Trust, and Trade-Offs report, released 16 July 2025, presents a survey rather than a product review. NORC at the University of Chicago fielded the questionnaire from 30 April to 14 May 2025 among 1,060 nationally representative U.S. teens aged 13 to 17, with an overall margin of sampling error of plus or minus 4.2 percentage points. Respondents were given a specific definition of AI companions as apps 'designed to have conversations that feel personal and meaningful,' distinct from tools that 'just answer questions,' and were told the survey excluded homework helpers, image generators, or voice assistants. Of the full sample, 758 teens reported having used an AI companion.

What the evidence says

The report's own figures state that 72% of teens have used an AI companion at least once, 52% qualify as regular users interacting at least a few times a month, and 13% use one daily (8% several times daily, 5% once daily). Among the 758 users, 33% reported choosing to discuss something important or serious with an AI companion instead of a real person, 34% reported feeling uncomfortable with something a companion said or did, and 24% reported sharing personal or private information such as their name or location. The same base reported spending more time with real friends than companions by a wide margin: 80% said more time with friends, against 6% who spend more time with companions. These are self-reported figures from a single fielding period, not an observed-behavior measure.

What it asks of people

The survey does not evaluate what any single companion product asks of a user; it measures how teens describe their own use across the category, naming Character.AI, Replika, Nomi, and CHAI as examples of the products covered. The report's authors state that platforms in this category 'rely on ineffective self-reporting for age assurance,' a description of industry practice generally rather than a finding against one named company's specific verification system, which this entry does not independently verify.

Privacy and safeguards

Common Sense Media's own press release states the findings 'underscore the risks identified' in the organisation's separate risk assessment of companion platforms, and quotes the organisation backing 'first-in-the-nation legislation to ban these products for minors,' an advocacy position attributed here to Common Sense Media rather than presented as this entry's own recommendation. The survey itself does not test any platform's safety features directly; it reports what surveyed teens said about their comfort, disclosure, and trust.

• Does a cited statistic use this survey's full-sample base or its companion-users-only base, since the two produce different percentages?
• Does a company's age-assurance claim amount to more than the self-reported check this report describes as standard across the category?
• Is an advocacy recommendation, such as banning use under 18, being read as a settled finding rather than the organisation's stated position?

The survey's headline number, that most teens have tried an AI companion, is now a documented, dated data point rather than an impression; it does not measure why any single platform's design choices produced that adoption, a question the report's own authors leave for further research.